Datasets:

ccdv
/

arxiv-summarization

Dataset card Data Studio Files Files and versions Community

Subset (2)

section · 216k rows

Split (3)

train · 203k rows

article stringlengths 0 745k	abstract stringlengths 0 92.7k
a well - known consequence of the ergodic decomposition theorem is that the space of invariant probability measures of a topological dynamical system , endowed with the weak@xmath0 topology , is a metrizable choquet simplex .. ] the purpose of this paper is to show that every non - empty metrizable choquet simplex arises in this way within the _ logistic family _ of maps @xmath4}$ ] , where for each parameter @xmath5 in @xmath6 $ ] the _ logistic map _ @xmath7 \to [ 0 , 1]$ ] is defined by , @xmath8 to make a more precise statement note that @xmath9 is the unique point in @xmath10 $ ] at which the derivative of @xmath11 vanishes . we call @xmath9 the _ critical point _ of @xmath12 , and its @xmath1-limit set is called the _ post - critical set _ of @xmath11 . it is a compact set that is forward invariant by @xmath11 . the following is our main result . recall that for a compact topological space @xmath13 a continuous map @xmath14 is _ minimal _ , if every forward orbit of @xmath15 is dense in @xmath13 . * main theorem . * _ for each non - empty metrizable choquet simplex @xmath16 there is a parameter @xmath17 $ ] such that the post - critical set of the logistic map @xmath11 is a cantor set , the restriction of @xmath11 to this set is minimal , and such that the space of invariant probability measures supported by this set , endowed with weak@xmath0 topology , is affine homeomorphic to @xmath16 . _ the first result of this kind was shown by downarowicz in @xcite , who showed that every non - empty metrizable choquet simplex arises , up to an affine homeomorphism , as the space of invariant probability measures of a `` minimal cantor system '' ; that is , a dynamical system generated by a minimal homeomorphism of a cantor set . in fact he showed that the minimal cantor system can be taken as a `` 0 - 1 toeplitz flow '' : a special type of subshift of @xmath18 . see also @xcite for a different approach to this result , and see @xcite for an analogous result in the case of actions of @xmath19 . the following corollary is a direct consequence of the main theorem and of the fact that for each non - empty polish space @xmath20 there is a metrizable choquet simplex whose set of extreme points is homeomorphic to @xmath20 , see for example @xcite . [ c : extreme realization ] for each non - empty polish space @xmath20 there is a parameter @xmath17 $ ] such that the post - critical set of the logistic map @xmath11 is a cantor set , the restriction of @xmath11 to this set is minimal , and such that the space of ergodic and invariant probability measures supported by this set , endowed with weak@xmath0 topology , is homeomorphic to @xmath20 . the special case where the polish space @xmath20 is compact and totally disconnected is precisely ( * ? * main theorem ) . the first result in this direction was shown by bruin , who gave an example of a parameter @xmath17 $ ] such that the post - critical set of @xmath12 is a cantor set where @xmath12 is minimal , but not uniquely ergodic ( * ? ? ? * theorem 4 ) . the proof of the main theorem is based on the tools developed by bruin in @xcite , and by bruin , keller and st . pierre in @xcite . one of the interesting features of the main theorem , in constrast with the other realization results mentioned above , is that the systems we consider have a natural differentiable structure . it turns out that , for the parameters @xmath17 $ ] given by ( the proof of ) the main theorem , the invariant measures supported by the post - critical set of @xmath12 correspond precisely to those invariant measures @xmath21 of @xmath12 whose lyapunov exponent @xmath22 vanishes ( * ? ? ? * lemma 21 ) . it also turns out that every invariant probability measure supported on the post - critical set of @xmath12 is an `` equilibrium state of @xmath12 for the potential @xmath23 '' . that is , if for each invariant measure @xmath21 we denote by @xmath24 its measure theoretic entropy , then the supremum @xmath25 is attained at each invariant probability measure supported by the post - critical set of @xmath12 , see ( * ? ? ? * lemma 21 ) . we thus obtain the following corollary of the main theorem . [ c : ergodic theory interval ] for each non - empty metrizable choquet simplex @xmath16 there is a parameter @xmath17 $ ] verifying the conclusions of the main theorem , and such that in addition the space of invariant probability measures of @xmath12 ( resp . equilibrium states of @xmath12 for the potential @xmath26 ) that are of zero lyapunov exponent , endowed with the weak@xmath0 topology , is affine homeomorphic to @xmath16 . this result is in sharp contrast with the fact that for a logistic map there can be at most one ergodic equilibrium state whose lyapunov exponent is strictly positive . for future reference we state an holomorphic version of corollary [ c : ergodic theory interval ] , shown in appendix [ a : ergodic theory ] . for a complex parameter @xmath27 denote by @xmath28 the quadratic polynomial defined by @xmath29 viewed as a dynamical system acting on @xmath30 . [ c : ergodic theory holomorphic ] for each non - empty metrizable choquet simplex @xmath16 there is a parameter @xmath17 $ ] verifying the conclusions of the main theorem , and such that in addition , if we denote by @xmath31 the hausdorff dimension of the julia set of @xmath28 , then the space of invariant probability measures of @xmath28 ( resp . equilibrium states of @xmath28 for the potential @xmath32 ) that are of zero lyapunov exponent , endowed with the weak@xmath0 topology , is affine homeomorphic to @xmath16 . we end this introduction by stating some questions that arise naturally from the main theorem . to do this , for each @xmath17 $ ] we will denote by @xmath33 the post - critical set of @xmath12 . given a compact metrizable topological space @xmath13 and a continuous map @xmath14 , the main theorem implies that there is a parameter @xmath17 $ ] and an affine homeomorphism @xmath34 between the space of invariant probability measures supported on @xmath33 . it is thus natural to ask whether the parameter @xmath17 $ ] and @xmath34 can be chosen in such a way that @xmath34 is induced by a continuous map between @xmath33 and @xmath13 . more precisely , the question is if @xmath17 $ ] and @xmath34 can be chosen in such a way that there is a continuous map @xmath35 such that for each invariant probability measure @xmath21 supported by @xmath33 we have @xmath36 . this type of problem is very well understood in the setting of minimal cantor systems : giordano , putnam , and skau have shown in @xcite that for two minimal cantor systems @xmath37 and @xmath38 there exists a homeomorphism @xmath39 that induces an affine homeomorphism between the corresponding spaces of invariant probability measures , if , and only if , @xmath37 and @xmath40 are `` orbit equivalent '' : there is a homeomorphism between @xmath13 and @xmath41 mapping each orbit of @xmath15 to an orbit of @xmath42 . furthermore , to each minimal cantor system @xmath37 one can associate a dimension group that is a complete invariant for the orbit equivalence relation @xcite : two minimal cantor systems are orbit equivalent if , and only if , the corresponding dimension groups are isomorphic as ordered groups with unit . and taking values in @xmath43 , by the subgroup of functions whose integral with respect to each invariant measure vanishes ; the positive cone is the set of those classes containing a function taking values in @xmath44 , and the unit is the class of the constant function equal to 1 . see for example ( * ? ? ? * theorem 1.13 ) . ] it is thus natural to look for a special class of minimal cantor systems realizing all of the orbit equivalent classes . since the dimension group associated to each toeplitz flow contains the dimension group of an odometer as a subgroup ( * ? ? ? * section 4.1 ) , the class of toeplitz flows is not sufficient to realize all orbit equivalence classes , in spite of the fact that this class realizes all the non - empty metrizable choquet simplices as sets of invariant probability measures @xcite . so the question remains whether minimal post - critical sets of logistic maps realize all orbit equivalence classes . in order to formulate a precise question we will consider natural extensions to stay in the class of minimal cantor systems , and use the generalized odometer associated to a kneading map , see [ ss : unimodal ] , [ ss : kneading odometer ] for definitions . does every orbit equivalence class contain the natural extension of a generalized odometer associated to a kneading map ? does every uniquely ergodic orbit equivalence class contain the natural extension of a generalized odometer associated to a kneading map ? it is well - known that every odometer can be realized , up to a homeomorphism , as the post - critical set of an infinitely renormalizable logistic map , see also @xcite . in [ ss : example ] we give an example of a uniquely ergodic generalized odometer associated to a kneading map , whose natural extension is not orbit equivalent to an odometer , nor to a toeplitz flow . although corollary [ c : extreme realization ] is stronger than ( * ? ? * main theorem ) , we use this last result in the proof of the main theorem to deal with case of finite dimensional choquet simplices . we have stated the main theorem and corollary [ c : extreme realization ] for the logistic family for simplicity . we show that an analogous statement holds for each full family of unimodal maps , as well as for the family of symmetric tent maps . see [ ss : unimodal ] for definitions . in fact , for each infinitely dimensional metrizable choquet simplex we construct kneading map @xmath45 such that the conclusions of the main theorem hold for each unimodal map whose kneading map is @xmath45 , see [ s : doubly resonant kneading ] . furthermore the kneading map satisfies for every @xmath46 the inequality @xmath47 ( part 1 of lemma [ l : doubly resonant kneading ] ) , and therefore every full family of unimodal maps , as well as the family of symmetric tent maps , contains a unimodal map whose kneading map is @xmath45 . for the case of finite dimensional choquet simplices see ( * ? ? ? * remark 1 ) . similarly , it follows from ( * ? ? ? * lemma 21 ) that corollary [ c : ergodic theory interval ] holds for every full family of -unimodal maps . see @xcite for a survey on realization results concerning toeplitz flows . see @xcite for the realization of some concrete simplices as the space of invariant measures of minimal cantor systems . in this section we explain the strategy of the proof of the main theorem and simultaneously describe the organization of the paper . we only deal with infinitely dimensional choquet simplices , the finite dimensional case being covered by ( * ? ? ? * main theorem ) . we use a result of lazar and lindenstrauss that characterizes infinite dimensional metrizable choquet simplices as inverse limits of stochastic matrices , see theorem [ t : choquet as inverse limit ] in [ ss : choquet simplices ] . we describe the logistic maps in the main theorem through their associated `` kneading map '' , see [ ss : unimodal ] for the definition of kneading map and further background on unimodal maps . in fact , the conclusions of the main theorem are valid for each unimodal map having the same kneading map as @xmath12 . to ensure that the post - critical set is a cantor set where the unimodal map is minimal , it is enough to require that the kneading map diverges to @xmath48 ( proposition [ p : continuous unimodal ] ) . in [ ss : doubly resonant kneading ] we introduce a class of kneading maps that diverge to @xmath49 . in [ ss : proof of main theorem ] we state a result describing the space of invariant measures supported on the post - critical set of a unimodal map with a kneading map in this class ( theorem [ t : invariant of doubly resonant ] ) . in [ ss : proof of main theorem ] we also give a proof of the main theorem assuming theorem [ t : invariant of doubly resonant ] . in [ ss : kneading odometer ] we recall the definition and some properties of the generalized odometer associated to a kneading map , that was introduced in @xcite . in [ ss : measure isomorphism ] we show that for a unimodal map whose kneading map @xmath45 is as in [ ss : doubly resonant kneading ] , the space of invariant probability measures supported by the post - critical set is affine homeomorphic to that of the generalized odometer associated to @xmath45 . in turn , this space is affine homeomorphic to the corresponding space of the bratteli - vershik system associated to @xmath45 , introduced in @xcite ; see ( * ? ? ? * proposition 2 ) or theorem [ t : reduction to bratteli ] in [ ss : kneading bratteli - vershik ] . the advantage of this last space is that it can be described explicitly as an inverse limit of some `` transition matrices '' , see [ ss : bratteli - vershik ] . we calculate the transition matrices and some of their products in [ ss : transition matrices ] , and give the proof of theorem [ t : invariant of doubly resonant ] in [ ss : proof of invariant of doubly resonant ] . in [ ss : example ] we give an example of a uniquely ergodic generalized odometer associated to a kneading map whose natural extension is not orbit equivalent to a toeplitz flow . in appendix [ a : ergodic theory ] we give the proof of corollary [ c : ergodic theory holomorphic ] . we are grateful with christian skau for his help with dimension groups of toeplitz flows , and an anonymous referee for his / her suggestions to improve the presentation of the paper . after fixing some notation in [ ss : linear algebra ] , we review some concepts and results about choquet simplices ( [ ss : choquet simplices ] ) , and unimodal maps ( [ ss : unimodal ] ) . throughout the rest of this article @xmath50 denotes the ring of strictly positive integers and @xmath51 . we will use the interval notation for subsets of @xmath44 : for @xmath52 such that @xmath53 , we put @xmath54 { : = } \ { k \in { \mathbb{n}}_0 \mid n \le k \le n ' \},\ ] ] and we put @xmath55 { : = } \emptyset$ ] when @xmath56 . given a non - empty finite set @xmath57 , for each @xmath58 we denote by @xmath59 the vector having all of its coordinates equal to @xmath60 , except for the coordinate corresponding to @xmath61 that is equal to @xmath62 . notice in particular that @xmath63 is a base of @xmath64 . furthermore we will denote by @xmath65 the _ unit simplex _ in @xmath64 , which is defined as the ( closed ) convex hull of @xmath63 in @xmath64 , and by @xmath66 the norm on @xmath64 defined by @xmath67 . observe that @xmath66 is constant equal to 1 on @xmath68 . given non - empty finite sets @xmath69 denote by @xmath70 the group of matrices whose entries are real and indexed by @xmath71 . for a matrix @xmath72 we denote by @xmath73 the transpose of @xmath74 , and for @xmath75 we denote by @xmath76 the corresponding entry of @xmath74 , and by @xmath77 the corresponding column vector of @xmath74 . given column vectors @xmath78 in @xmath64 we denote by @xmath79 the matrix in @xmath70 whose column vector corresponding to the coordinate @xmath80 is equal to @xmath81 . we say that a matrix @xmath74 is ( left ) _ stochastic _ if all of its entries are non - negative and if the sum of all the entries in each column is equal to 1 . observe that a stochastic matrix in @xmath70 maps @xmath82 into @xmath68 , and that the product of stochastic matrices is stochastic . [ l : non expanding ] let @xmath69 be non - empty finite sets and let @xmath72 be a stochastic matrix . then for each @xmath83 we have @xmath84 putting @xmath85 and @xmath86 , we have @xmath87 a compact , convex , and metrizable subset @xmath16 of a locally convex real vector space is said to be a ( metrizable ) _ choquet simplex _ , if for each @xmath88 there is a unique probability measure @xmath21 that is supported on the set of extreme points of @xmath16 , and such that @xmath89 . see for example @xcite for several characterizations of choquet simplices . in the proof of the main theorem we will make use of the following characterization of infinite dimensional metrizable choquet simplices . [ t : choquet as inverse limit ] given an infinite dimensional choquet simplex @xmath16 , for each @xmath90 there is a surjective affine map @xmath91 } \to \delta_{[0 , n]}$ ] such that @xmath92 } , a_n)$ ] is affine homeomorphic to @xmath16 . the following lemma is a simple consequence of the previous theorem . we will say that a sequence @xmath93 as in the theorem is _ normalized _ , if for every @xmath90 and @xmath94 $ ] we have @xmath95 . [ l : normalized choquet ] for every infinite dimensional metrizable choquet simplex @xmath16 there is a normalized sequence of linear maps @xmath93 such that @xmath96 } , a_n)$ ] is affine homeomorphic to @xmath16 . let @xmath97 be a sequence of affine maps given by theorem [ t : choquet as inverse limit ] . for each @xmath90 define a permutation @xmath98 of @xmath99 $ ] by induction as follows . let @xmath100 be the identity , and suppose that for some @xmath90 the permutation @xmath98 is already defined . since @xmath101 maps @xmath102}$ ] surjectively onto @xmath103}$ ] , for each @xmath94 $ ] there is @xmath104 $ ] such that @xmath105 . equivalently , there is a map @xmath106 \to [ 0 , n + 1]$ ] such that for each @xmath94 $ ] we have @xmath107 . let @xmath108 be the unique permutation of @xmath109 $ ] such that for each @xmath94 $ ] we have @xmath110 . for each @xmath111 let @xmath112 } \to { \mathbb{r}}^{[0 , n]}$ ] be the linear map so that for each @xmath94 $ ] we have @xmath113 . then , by the definition of @xmath114 it follows that for each @xmath90 the linear map @xmath115 maps @xmath102}$ ] surjectively onto @xmath103}$ ] , and that for every @xmath94 $ ] we have @xmath116 . therefore the sequence of linear maps @xmath93 is normalized , and @xmath117 induces a linear homeomorphism between @xmath92 } , { \widetilde{a}}_n)$ ] and @xmath92 } , a_n)$ ] . we end this section with the following general lemma . [ l : equivalence ] for each @xmath90 let @xmath118 } \to { \mathbb{r}}^{[0 , n]}$ ] be stochastic matrices such that @xmath119 } \ } < + \infty.\ ] ] then the inverse limits @xmath120 } , a_n)$ ] and @xmath120 } , b_n)$ ] are affine homeomorphic . in particular , if both @xmath93 and @xmath121 are normalized and @xmath122 then the inverse limits @xmath120 } , a_n)$ ] and @xmath120 } , b_n)$ ] are affine homeomorphic . let @xmath123},a_n).$ ] for each @xmath124 such that @xmath125 define @xmath126}.\ ] ] in particular we have @xmath127 . when @xmath128 we have @xmath129 from lemma [ l : non expanding ] we get @xmath130 and since @xmath131 } : \\|a_nv - b_nv\\|_1\},\ ] ] we deduce , after an induction argument , that @xmath132 } \}.\end{aligned}\ ] ] by hypothesis and from equation ( [ eq3 ] ) , we deduce that for a fixed @xmath90 , sequence @xmath133 is a cauchy sequence in @xmath134}$ ] . we denote by @xmath135 its limit . observe that for each @xmath90 , the sequence @xmath136 converges to both @xmath137 and @xmath135 . this implies that @xmath138},b_n).$ ] thus the transformation @xmath139},a_n ) \to \varprojlim_{n}(\delta_{[0,n+1]},b_n)\ ] ] given by @xmath140 is well defined . this map is clearly affine . we will show that @xmath34 is a homeomorphism . in order to verify that @xmath34 is continuous , we just need to show that for each @xmath141 the map @xmath142 is continuous . as @xmath92 } , a_n)$ ] is metrizable we just need to show that @xmath142 is sequentially continuous . to do this , fix @xmath143 and consider a sequence @xmath144 in @xmath145},a_n)$ ] that converges to @xmath146 . for each @xmath147 such that @xmath148 we have @xmath149 we choose @xmath150 sufficiently large so that @xmath151 and @xmath111 such that @xmath152 . we get @xmath153 this shows the continuity of @xmath142 , and hence that of @xmath34 . to show that @xmath34 is a homeomorphism we define in a similar way , @xmath154},b_n ) \to \varprojlim_{n}(\delta_{[0,n+1]},a_n).\ ] ] observe that by lemma [ l : non expanding ] we get @xmath155 by definition of @xmath156 , we have @xmath157 and from we have @xmath158 . thus we conclude that @xmath156 and @xmath34 are inverse of each other . a continuous map @xmath159\to [ 0,1]$ ] is _ unimodal _ if @xmath160 , and if there exists a point @xmath161 $ ] such that @xmath2 is strictly increasing on @xmath162 $ ] , and strictly decreasing on @xmath163 $ ] . the point @xmath164 is called the _ turning _ or _ critical point _ of @xmath2 . for each @xmath17 $ ] the logistic map @xmath12 is a unimodal with critical point @xmath9 . let @xmath2 be a unimodal map with critical point @xmath164 . the @xmath1-limit of @xmath164 will be called the _ post - critical set _ of @xmath2 . when either @xmath165 or @xmath166 , it is easy to see that the post - critical set of @xmath2 reduces to a single point . we will thus ( implicitly ) assume from now on that for each unimodal map @xmath2 that we consider we have @xmath167 . to describe the dynamics of a unimodal map @xmath2 on its post - critical set , we will make the following definitions . let @xmath164 be the critical point of @xmath2 and for each @xmath168 put @xmath169 . define the sequence of compact intervals @xmath170 inductively by @xmath171 $ ] , and for each @xmath172 , by @xmath173 & \text{otherwise}. \end{cases}\ ] ] an integer @xmath168 will be called a _ cutting time _ if @xmath174 . we will denote by @xmath175 the sequence of all cutting times . from our assumption that @xmath167 it follows that @xmath176 and @xmath177 . it can be shown that if @xmath178 and @xmath179 are consecutive cutting times , then @xmath180 is again a cutting time , and that this cutting time is less than or equal to @xmath178 when @xmath2 has no periodic attractors , see for example @xcite . that is , if @xmath2 has no periodic attractors then for each @xmath181 there is a non - negative integer @xmath182 , such that @xmath183 , and @xmath184 putting @xmath185 , the function @xmath186 so defined is called the _ kneading map _ of @xmath2 . it follows from the recursion formula above , and from @xmath176 , that the sequence @xmath175 of cutting times is determined by @xmath45 . we will say that a function @xmath186 is a _ kneading map _ if there is a unimodal map @xmath2 with critical point @xmath164 , such that @xmath167 , such that @xmath2 has no periodic attractors and such that the kneading map of @xmath2 is equal to @xmath45 . if we denote by @xmath187 the lexicographical ordering in @xmath188 , then a function @xmath186 is a kneading map if and only if @xmath185 , for each @xmath181 we have @xmath183 , and if for each @xmath181 we have @xmath189 see @xcite . notice in particular that , if @xmath186 is non - decreasing , @xmath185 and for each @xmath181 we have @xmath183 , then @xmath45 is a kneading map . we will need the following well known facts , see for example the proof of ( * ? ? ? * proposition 4 ) for precise references . [ p : continuous unimodal ] let @xmath2 be a unimodal map whose kneading map diverges to @xmath48 . then the post - critical set of @xmath2 is a cantor set , and the restriction of @xmath2 to this set is minimal and has zero topological entropy . furthermore , if @xmath190 is a unimodal map having the same kneading map as @xmath2 , then the space of invariant probability measures of @xmath190 supported on the post - critical set of @xmath190 is affine homeomorphic to that of @xmath2 . the purpose of this section is to prove the main theorem assuming a result we state as theorem [ t : invariant of doubly resonant ] . we start introducing a class of kneading maps in [ ss : doubly resonant kneading ] . in [ ss : proof of main theorem ] we state a result describing the space of invariant probability measures supported on the post - critical set of a unimodal map with a kneading map in this class ( theorem [ t : invariant of doubly resonant ] ) . in [ ss : proof of main theorem ] we also give a proof of the main theorem assuming theorem [ t : invariant of doubly resonant ] . throughout the rest of this paper we denote by @xmath191 the sequence of integers defined by @xmath192 . note that for each @xmath90 we have @xmath193 . for each @xmath90 let @xmath194}$ ] be given and let @xmath195 be an increasing sequence of integers such that @xmath196 , and such that for each @xmath90 we have , @xmath197 if we put @xmath198 , then we will define a kneading map @xmath199 as follows . for each @xmath200 put @xmath201 \text { and } { j}_n = [ q_{r_{n } } + 1 , q_{r_{n + 1 } - 1}],\ ] ] and note that @xmath202 is a partition of @xmath50 . furthermore , for each @xmath90 and @xmath203 $ ] define @xmath204\ ] ] and @xmath205.\ ] ] by the collection @xmath206 \}$ ] is a partition of @xmath207 . since for each @xmath90 we have @xmath208 , the collection @xmath209 \}$ ] is a partition of @xmath210 . with these notations we put @xmath211 note that @xmath212 $ ] , @xmath213 and that for each @xmath90 we have @xmath214 [ l : doubly resonant kneading ] for @xmath198 , @xmath215 and @xmath216 as above , the following properties hold . 1 . for each @xmath46 we have @xmath217 . the function @xmath216 is a kneading map and for every @xmath218 we have @xmath219 put @xmath220 . * * for @xmath221 $ ] we have @xmath222 so the inequality is satisfied in this case . let @xmath223 be such that @xmath224 , so there is @xmath90 such that @xmath225 $ ] . thus @xmath226 , so when @xmath227 we have @xmath228 . finally observe that @xmath229 * 2 . * let @xmath223 be such that @xmath230 , so there is @xmath172 such that @xmath231 . then @xmath232 , @xmath233 and therefore we have @xmath234 and @xmath235 in view of part 1 and the previous inequality , to show that @xmath45 is admissible we just need to show that for each @xmath236 $ ] and @xmath237 $ ] we have @xmath238 with strict inequality when @xmath239 . in fact , for each @xmath236 $ ] we have @xmath240 , so for each @xmath237 $ ] , we have @xmath241 and is satisfied . when @xmath239 we have @xmath242 , so inequality is strict in this case . the purpose of this section is to give a proof of the main theorem assuming the following one . [ t : invariant of doubly resonant ] for each @xmath90 let @xmath243}$ ] be given and put @xmath198 . furthermore , let @xmath244 be a strictly increasing sequence of integers such that @xmath196 and such that for each @xmath90 we have @xmath245 and let @xmath216 be the corresponding kneading map . define @xmath246 recursively by @xmath176 and @xmath247 , and assume that @xmath248 moreover , for each @xmath90 let @xmath249 } \to { \mathbb{r}}^{[0 , n]}$ ] be the stochastic matrix such that for each @xmath203 $ ] we have @xmath250 and such that @xmath251 then for each unimodal map @xmath2 whose kneading map is equal to @xmath216 , the post - critical set of @xmath2 is a cantor set , @xmath2 is minimal on this set , and the space of invariant probability measures of @xmath2 supported on this set is affine homeomorphic to @xmath92 } , \xi_n)$ ] . to prove the main theorem , we first remark that the case where the metrizable choquet simplex @xmath16 is finite dimensional is given by ( * ? ? ? * main theorem ) . so from now on we assume that @xmath16 is infinite dimensional . then by lemma [ l : normalized choquet ] there is a sequence of normalized affine maps @xmath93 such that @xmath92 } , a_n)$ ] is affine homeomorphic to @xmath16 . in view of lemma [ l : equivalence ] , we just need to find @xmath252 and @xmath253 as in the statement of theorem [ t : invariant of doubly resonant ] , for which is satisfied and such that @xmath254 this is shown in the following lemma , thus completing the proof of the main theorem . [ l : approximation ] for each @xmath90 let @xmath255}$ ] be given . then there are @xmath252 and @xmath253 as in the statement of theorem [ t : invariant of doubly resonant ] for which is satisfied , and such that @xmath256 given @xmath90 and a non - zero vector @xmath257}$ ] with non - negative coordinates , we will denote by @xmath258 the unique vector in @xmath103}$ ] proportional to @xmath259 . we will define @xmath198 and @xmath244 by induction as follows . put @xmath196 , fix @xmath260 , and assume that for some @xmath90 the numbers @xmath261 and the vectors @xmath262 are already defined , in such a way that for each @xmath263 $ ] we have @xmath264 for each @xmath265 $ ] let @xmath266 be defined in such a way that @xmath267 note that these choices determine @xmath268 . * we will show now that for each @xmath265 $ ] we have @xmath269 using the recursion formula @xmath270 and @xmath271 , we get by induction that for every @xmath272 $ ] such that @xmath273 , we have @xmath274 . in particular for every @xmath275 $ ] we have @xmath276 since @xmath277 , it follows by induction that for every @xmath275 $ ] we have @xmath278 hence for each @xmath265 $ ] we have @xmath279 as wanted . * we will show that we can choose @xmath194}$ ] is such a way that @xmath280 for @xmath281 we denote by @xmath282 $ ] integer part of @xmath283 . put @xmath284 , @xmath285 , @xmath286 , for each @xmath94 $ ] put @xmath287 + 1 ) , \text { and put } \zeta = \zeta_1 + \cdots + \zeta_n.\ ] ] since for each @xmath94 $ ] we have @xmath288 , we obtain @xmath289 . so for each @xmath290 $ ] we have @xmath291 this shows that the vector @xmath292}$ ] satisfies @xmath293 . thus , if for each @xmath294 $ ] we put @xmath295 , and if we put @xmath296 , then is satisfies for this choice of @xmath297 . it remains to show that each of the coordinates of @xmath297 belongs to @xmath50 . by definition for each @xmath94 $ ] the integer @xmath298 is a strictly positive multiple of @xmath299 , so the coordinates of @xmath297 are integers and for each @xmath294 $ ] we have @xmath300 . finally observe that by with @xmath301 we have , @xmath302 let @xmath297 be given by part 2 and put @xmath303 . this completes the inductive definition of @xmath253 and @xmath252 . to finish the proof of the lemma just observe that the inequalities imply , and the inequalities imply . the purpose of this section is to recall the definition of the generalized odometer and the bratteli - vershik system associated to a kneading map , that were introduced in @xcite and @xcite , respectively . we start recalling the definition of the generalized odometer in [ ss : kneading odometer ] . after briefly recalling the concepts of bratteli diagram ( [ ss : diagram ] ) and bratteli - vershik system ( [ ss : bratteli - vershik ] ) , we define the bratteli - vershik system associated to a kneading map in [ ss : kneading bratteli - vershik ] . see for example @xcite for background on generalized odometers , and @xcite and references therein for background and further properties of bratteli - vershik systems . let @xmath186 be a kneading map and put @xmath304 if we denote by @xmath175 the strictly increasing sequence of positive integers defined recursively by @xmath176 and @xmath305 , it can be shown that for each non - negative integer @xmath111 there is a unique sequence @xmath306 in @xmath307 , that has at most finitely many @xmath62 s , and such that @xmath308 . the sequence @xmath309 is also characterized as the unique sequence in @xmath310 with finitely many @xmath62 s such that @xmath308 , and that it is minimal with this property with respect to the lexicographical order in @xmath310 . when @xmath45 diverges to @xmath48 the map defined on the subset @xmath311 of @xmath307 by @xmath312 , extends continuously to a map @xmath313 which is onto , minimal , and such that @xmath314 is well defined on @xmath315 ; see ( * ? ? ? * lemma 2 ) . we call @xmath316 the _ _ generalized odometer _ _ is such that for each @xmath46 we have @xmath317 . ] associated to @xmath45 . given @xmath318 and an integer @xmath319 , put @xmath320 . observe that @xmath321 is non - decreasing with @xmath111 , and when @xmath283 has infinitely many @xmath62 s , @xmath322 as @xmath323 . on the other hand , if @xmath283 has at most a finite number of @xmath62 s , then @xmath324 is finite and @xmath325 . for @xmath326 different from @xmath327 we denote by @xmath328 the least integer such that @xmath329 . in ( * theorem 1 ) it is shown that if @xmath17 $ ] is a parameter such that the kneading map of the logistic map @xmath12 is equal to @xmath45 , then for each @xmath330 with infinitely many @xmath62 s the sequence of intervals @xmath331 is nested and that @xmath332 is reduced to a point belonging to the post - critical set @xmath333 of @xmath12 . furthermore , if we denote this point by @xmath334 and for @xmath319 we put @xmath335 , then the map @xmath336 so defined is continuous and conjugates the action of @xmath337 on @xmath307 , to the action of @xmath12 on @xmath333 . a _ bratteli diagram _ is an infinite directed graph @xmath338 , such that the vertex set @xmath57 and the edge set @xmath339 can be partitioned into finite sets @xmath340 with the following properties : * @xmath341 is a singleton . * for every @xmath342 , each edge in @xmath343 starts in a vertex in @xmath344 and arrives to a vertex in @xmath345 . * all vertices in @xmath57 have at least one edge starting from it , and all vertices except @xmath346 have at least one edge arriving to it . for a vertex @xmath347 we will denote by @xmath348 the vertex where @xmath349 starts and by @xmath350 the vertex to which @xmath349 arrives . a _ path _ in @xmath351 is by definition a finite ( resp . infinite ) sequence @xmath352 ( resp . @xmath353 ) such that for each @xmath354 ( resp . @xmath355 ) we have @xmath356 . note that for each vertex @xmath61 distinct from @xmath346 there is at least one path starting at @xmath346 and arriving to @xmath61 . an _ ordered bratteli diagram _ @xmath357 is a bratteli diagram @xmath338 together with a partial order @xmath358 on @xmath339 , so that two edges are comparable if and only if they arrive at the same vertex . for each @xmath342 and @xmath359 the partial order @xmath360 induces an order on the set of paths from @xmath346 to @xmath57 as follows : @xmath361 if and only there exists @xmath362 such that @xmath363 and such that for each @xmath364 we have @xmath365 . we will say that an edge @xmath349 is _ maximal _ ( resp . _ minimal _ ) if it is maximal ( resp . minimal ) with respect to the order @xmath360 on the set of all edges in @xmath339 arriving at @xmath350 . note that for each vertex @xmath61 distinct from @xmath346 there is precisely one path starting at @xmath346 and arriving to @xmath61 that is maximal ( resp . minimal ) with respect to the order @xmath360 . it is characterized as the unique path starting at @xmath346 and arriving at @xmath61 consisting of maximal ( resp . minimal ) edges . fix an ordered bratteli diagram @xmath366 . we denote by @xmath367 set of all infinite paths in @xmath368 starting at @xmath346 . for a finite path @xmath369 starting at @xmath346 we denote by @xmath370 the subset of @xmath367 of all infinite paths @xmath371 such that for all @xmath372 we have @xmath373 . we endow @xmath367 with the topology generated by the sets @xmath370 . then each of this sets is clopen , so @xmath367 becomes a compact hausdorff space with a countable basis of clopen sets . we will denote by @xmath374 ( resp . @xmath375 ) the set of all elements @xmath376 of @xmath367 so that for each @xmath342 the edge @xmath377 is a maximal ( resp . minimal ) . it is easy to see that each of these sets is non - empty . from now on we assume that the set @xmath375 is reduced to a unique point , that we will denote by @xmath378 . we will then define the transformation @xmath379 as follows : * @xmath380 . * given @xmath381 , let @xmath342 be the smallest integer such that @xmath377 is not maximal . then we denote by @xmath382 the successor of @xmath377 and by @xmath383 the unique minimal path starting at @xmath346 and arriving to @xmath384 . then we put , @xmath385 the map @xmath386 is continuous , onto and invertible except at @xmath378 . for @xmath342 and @xmath359 we denote by @xmath387 the number of paths starting at @xmath346 and arriving to @xmath61 , and put @xmath388 . let @xmath389 be the matrix such that for each @xmath390 and @xmath391 the entry @xmath392 is equal to the number of edges starting at @xmath61 and arriving to @xmath80 . observe that @xmath393 , so if we put @xmath394 and for each @xmath342 we denote by @xmath395 the diagonal matrix defined by @xmath396 , then the matrix @xmath397 is stochastic . the following result is well - known , see ( * ? ? ? * lemma 14 ) for a proof in the precise setting considered here . recall that for a finite set @xmath57 we denote by @xmath68 the unit simplex in @xmath64 . [ l : medidas bratteli ] the space of probability measures on @xmath367 that are invariant by @xmath386 , endowed with the weak@xmath0 topology , is affine homeomorphic to @xmath398 . given a kneading map @xmath45 we will now define an ordered bratteli diagram @xmath399 that was introduced by bruin in @xcite . we start defining the bratteli diagram @xmath351 : * @xmath400 , @xmath401 and for @xmath402 , @xmath403 * for @xmath342 , @xmath404 note that for every @xmath402 , each vertex in @xmath405 different from @xmath406 has at most one edge arriving at it . besides @xmath407 , the only edge that can arrive to @xmath408 is @xmath409 , that only exists when @xmath410 . so to define the partial order @xmath360 , we just have to define it , for each @xmath402 , between @xmath411 and @xmath412 when both exist : we put @xmath413 . the rest of the edges are maximal and minimal at the same time . note that for @xmath181 the set @xmath414 is reduced to a point if and only if @xmath415 . so , if for each large @xmath181 we have @xmath415 , then the set @xmath416 is finite . otherwise , it follows that the set @xmath416 is a cantor set . it is straight forward to check that the infinite path @xmath417 is the unique minimal path in @xmath418 . therefore there is a well defined map @xmath419 , see [ ss : bratteli - vershik ] . the following is ( * ? ? * proposition 2 ) , and the last statement follows from ( * ? ? ? * lemma 2 ) . [ t : reduction to bratteli ] let @xmath45 be a kneading map that diverges to @xmath48 , and consider the corresponding bratteli - vershik system @xmath420 and generalized odometer @xmath316 . then there is a homeomorphism between @xmath416 and @xmath307 that conjugates the action of @xmath421 on @xmath416 to the action of @xmath337 on @xmath307 . in particular @xmath420 is minimal . we will also need the following lemma . [ l : recurrence on paths ] let @xmath45 be a kneading map such that for every @xmath46 we have @xmath47 , and such that @xmath422 as @xmath423 . let @xmath424 be the sequence defined recursively by @xmath176 and @xmath305 . then for every @xmath425 we have @xmath426 , @xmath427 , and for every @xmath428 we have @xmath429 . that @xmath430 is a direct consequence of the definition of @xmath405 and the hypothesis that @xmath431 . when @xmath432 , we have @xmath433 and for all @xmath434 we have @xmath435 . so the assertions are satisfied in this case . suppose by induction that the assertions of the lemma hold for some @xmath342 . then by we have , @xmath436 on the other hand , for each @xmath437 contained in @xmath405 we have @xmath438 . finally , for each @xmath439 not in @xmath405 we have @xmath440 and @xmath441 this section is devoted to the proof of theorem [ t : invariant of doubly resonant ] . we start by showing that for each unimodal map whose kneading map is as in [ ss : doubly resonant kneading ] , the space of invariant probability measures supported on its post - critical set is affine homeomorphic to the space of invariant probability measures of the corresponding generalized odometer . in order to describe this space we calculate the transition matrices associated to the corresponding bratteli - vershik system . the key calculation of a suitable product of these transition matrices is stated as proposition [ p : full block ] in [ ss : transition matrices ] . the proof of theorem [ t : invariant of doubly resonant ] is given in [ ss : proof of invariant of doubly resonant ] . the purpose of this section is to prove the following proposition , whose proof is similar to that of ( * ? ? ? * theorem b ) . [ p : measure isomorphism ] let @xmath216 be a kneading map defined as in [ ss : doubly resonant kneading ] , and let @xmath442 be the corresponding generalized odometer . let @xmath2 be a unimodal map whose kneading map is equal to @xmath216 , and denote by @xmath443 its post - critical set . then the space of invariant probability measures of @xmath444 is affine homeomorphic to that of @xmath442 . the following lemma is similar to ( * ? ? ? * lemma 11 ) . observe that , since @xmath314 is well defined on @xmath445 , if we denote by @xmath446 the grand orbit of @xmath327 , then @xmath447 and all negative iterates of @xmath337 are well defined on @xmath448 . [ l : criterion satisfied ] let @xmath449 be a kneading map as in [ ss : doubly resonant kneading ] . let @xmath450 be the corresponding generalized odometer . then for each constant @xmath451 , and for every pair of distinct points @xmath452 in @xmath453 that are not in the grand orbit of @xmath327 , there is an integer @xmath454 satisfying @xmath455 let @xmath456 . * 1 . * as in the proof of ( * ? ? ? * lemma 10 ) , it can be shown that there is an integer @xmath457 such that @xmath458 * 2 . * let @xmath457 be the integer given by part 1 , and put @xmath459 and @xmath460 . assume without loss of generality that @xmath461 , so that @xmath462 . if @xmath463 then take @xmath464 . so we assume that @xmath465 . since @xmath466 we have @xmath467 , so there is @xmath90 such that @xmath468 and @xmath469 belong to @xmath470 . the definition of @xmath471 and of @xmath453 imply that for each @xmath472 $ ] we have @xmath473 . indeed , suppose by contradiction that for such a @xmath150 we have @xmath474 . then the definition of @xmath453 implies that for every @xmath475 $ ] we have @xmath476 . since @xmath477 , we get @xmath478 , which is a contradiction . therefore @xmath479 is such that for all @xmath480 $ ] we have @xmath481 . since @xmath482 is not in the grand orbit of @xmath327 this implies that @xmath483 . thus , from the definition of @xmath471 we have @xmath484 on the other hand , since @xmath461 , we have @xmath485 , so @xmath486 this shows that the integer @xmath487 satisfies the desired properties . since the logistic family is full there is a parameter @xmath17 $ ] such that the kneading map of the logistic map @xmath12 is @xmath216 . denote by @xmath488 the post - critical set of @xmath12 . by proposition [ p : continuous unimodal ] the spaces of invariant measures of @xmath489 and @xmath490 are affine homeomorphic . so , without loss of generality we assume that @xmath2 is a logistic map . this ensures the existence of the factor map @xmath491 defined above . since for every sufficiently large integer @xmath150 inequality is satisfied , ( * ? ? ? * lemma 11 ) implies there is a constant @xmath492 such that for every pair of distinct points @xmath493 in @xmath494 that are not in the grand orbit of @xmath327 and that satisfy @xmath495 we have @xmath496 . thus , from lemma [ l : criterion satisfied ] we deduce that @xmath497 is injective on @xmath498 . the rest of the proof follows as the proof of ( * ? ? ? * proposition 9 ) . for each @xmath90 let @xmath243}$ ] be given , and put @xmath198 . furthermore , let @xmath244 be a strictly increasing sequence of integers such that @xmath196 and such that for each @xmath90 we have @xmath245 and let @xmath220 be the corresponding kneading map defined in [ ss : doubly resonant kneading ] . let @xmath499 be the ordered bratteli - vershik diagram associated to the kneading map @xmath45 . from the definition of @xmath418 it follows that @xmath500 $ ] , and that for each @xmath501 $ ] we have @xmath502 $ ] . furthermore , for each @xmath200 , @xmath203 $ ] , and @xmath503 $ ] we have @xmath504 \cup \left ( 1 + \bigcup_{i = 0}^m \left ( { i}_{n + 1 , i } \cup { j}_{n + 1 , i } \right ) \right),\ ] ] and that when @xmath505 we have @xmath506 , and for @xmath507 $ ] we have @xmath508.\ ] ] note in particular that for every @xmath90 we have @xmath509 \setminus \left ( 1 + ( { i}_{n , n } \cup { j}_{n , n } ) \right),\ ] ] @xmath510.\ ] ] [ p : full block ] given @xmath90 , for each @xmath511 $ ] put @xmath512 then the columns of the matrix @xmath513 , [ q_{r_{n + 1 } } + 1 , q_{r_{n + 2 } - 1 } + 1]),\ ] ] are given by , @xmath514 and for every @xmath515 $ ] and @xmath516 such that @xmath517 , by @xmath518 in particular the rank of the matrix @xmath519 is equal to @xmath520 . the proof of this proposition depends on the following lemma . [ l : basic blocks ] fix @xmath90 and let @xmath515 $ ] . if @xmath203 $ ] then the set @xmath521 is equal to the disjoint union of @xmath522 $ ] and of @xmath523 , and we have @xmath524 ; \\ \vec{e}_{q_{r_n + m } + 1 } & \text{if } \ell - 1 \in q^{-1}(q_{r_n + m } ) . \end{cases}\end{gathered}\ ] ] when @xmath505 we have @xmath525 , @xmath526 $ ] , and @xmath527 it follows from the definition of @xmath418 that for each @xmath528 the set @xmath529 is equal to the disjoint union of @xmath530 and @xmath531 , and that @xmath532 on the other hand , for each @xmath533 $ ] we have @xmath534 and @xmath535 a direct computation using the fact that for every @xmath536 $ ] the set @xmath345 is the disjoint union of @xmath537 and @xmath538 $ ] , shows that @xmath539 . \end{cases}\end{gathered}\ ] ] fix @xmath90 and let @xmath515 $ ] . when @xmath203 $ ] the assertion of the lemma about @xmath521 follows from . on the other hand , by if @xmath540 and by if @xmath541 $ ] , we have @xmath542 \subset v_{q_{r_n + m } + 1}.\ ] ] then in this case the assertion of the lemma follows easily from with @xmath543 . suppose now that @xmath505 . by the definition of the sequence @xmath544 we have @xmath506 . the assertion about @xmath545 is given by with @xmath111 replaced by @xmath546 . since by we have @xmath547 \cap v_{q_{r_{n + 1 } - 1 } + 1 } = 1 + ( { i}_{n + 1 } \setminus { i}_{n + 1 , n + 1 } ) = \\ = [ q_{r_{n + 1 } -1 } + 2 , q_{r_{n + 1 } } + 1 - a_{n + 1 , n + 1}]\end{gathered}\ ] ] and since @xmath548 , we conclude from with @xmath549 , that @xmath550 on the other hand , for @xmath551 \cap v_{q_{r_{n + 1 } - 1 } + 1 } = 1 + \left ( { j}_{n + 1 } \setminus { j}_{n + 1 , n + 1 } \right),\ ] ] we have @xmath552 and that for @xmath551 \cap \left ( 1 + q^{-1}(q_{r_{n + 1 } - 1 } ) \right ) = 1 + { j}_{n + 1 , n + 1},\ ] ] we have @xmath553 this completes the proof of the lemma . fix @xmath90 . * we will show by induction that for each @xmath554 $ ] we have @xmath555 ; \\ \sum_{k = q_{r_n}}^{q_{r_n + m } } \vec{e}_{k + 1 } & \text{if } \ell - 1 \in { i}_{n + 1 , m } \cup { j}_{n + 1 , m } \text { and } m \in [ 0 , m_0 - 1 ] . \end{cases}\end{gathered}\ ] ] the case @xmath556 is given by with @xmath557 . suppose that this holds for some @xmath558 $ ] . observe that by the set @xmath559 is the disjoint union of @xmath560\ ] ] and @xmath561 in view of with @xmath562 we obtain @xmath563 for each @xmath564 $ ] we have @xmath565 \\ \sum_{k = q_{r_n}}^{q_{r_n + m } } \vec{e}_{k + 1 } & \text{if } \ell - 1 \in { i}_{n + 1 , m } \cup { j}_{n + 1 , m } \text { and } m \in [ 0 , m_0 - 1 ] . \end{cases}\end{gathered}\ ] ] finally , for each @xmath566 such that @xmath567 we have @xmath568 this completes the proof of the induction step . a direct computation using with @xmath569 and , gives @xmath570 and for every @xmath515 $ ] and @xmath516 such that @xmath571 , @xmath572 the assertion of the proposition is then a direct consequence of the definition of the matrices @xmath573 and lemma [ l : recurrence on paths ] . let @xmath2 be a unimodal having @xmath45 as kneading map . that the post - critical set of @xmath2 is a cantor set and that @xmath2 is minimal on this set is given by proposition [ p : continuous unimodal ] . in view of proposition [ p : measure isomorphism ] and theorem [ t : reduction to bratteli ] , it is enough to prove that the space of invariant probability measures of the bratteli - vershik system @xmath420 is affine homeomorphic to @xmath574 } , \xi_n)$ ] . for each @xmath90 let @xmath575}$ ] be the stochastic matrix defined by @xmath576 using the definition of @xmath577 in the statement of proposition [ p : full block ] , for each @xmath578 $ ] we put @xmath579 furthermore , when @xmath172 , we denote by @xmath580 } \to { \mathbb{r}}^{[0 , n]}$ ] the stochastic matrix defined for @xmath203 $ ] by @xmath581 and by @xmath582 a direct computation shows that for each @xmath172 we have @xmath583 therefore the sequence of maps @xmath584 define a continuous linear map @xmath585 } , a_n),\ ] ] mapping @xmath586 onto @xmath120 } , a_n)$ ] . by proposition [ p : full block ] the rank of the matrix @xmath587 is equal to @xmath546 , so @xmath588 is a homeomorphism and the inverse limits @xmath586 and @xmath120 } , a_n)$ ] are affine homeomorphic . in view of lemma [ l : equivalence ] , the following lemma together with the hypothesis @xmath589 imply that the inverse limit @xmath120 } , a_n)$ ] is affine homeomorphic to @xmath120 } , \xi_n)$ ] . this completes the proof of theorem [ t : invariant of doubly resonant ] . for each @xmath90 and @xmath590 $ ] we have @xmath591 furthermore @xmath592 , and @xmath593 by definition we have @xmath594 , and for each @xmath590 $ ] we have @xmath595 on the other hand , @xmath596 so the final assertion follows from . given @xmath597 put @xmath598 for each such @xmath599 we will construct a kneading map @xmath45 such that the dimension group associated to the generalized odometer @xmath316 , and hence to its natural extension , is isomorphic to @xmath600 . thus we deduce that every simple dimension group which is free of rank @xmath601 is isomorphic as ordered group to the dimension group associated to ( the natural extension of ) a generalized odometer associated to a kneading map . since the rational subdimension group of @xmath600 is @xmath602 , it follows from @xcite that the dimension group @xmath600 is not isomorphic to the dimension group associated to a toeplitz flow , nor to that of an odometer . before defining the kneading map @xmath45 , note that the dimension groups @xmath600 , @xmath603 , and @xmath604 are isomorphic to each other . so we can restrict to the case where @xmath605 . let @xmath606 be the integer determined by @xmath607 , and let @xmath608 $ ] be the continued fraction expansion of @xmath609 \setminus { \mathbb{q}}$ ] . consider the function @xmath186 defined by @xmath610 ; \\ k-1 & \mbox { if } l \in [ k + 1 , k + a_1 ] ; \\ k-1+\sum_{i=1}^na_i & \mbox { if } l \in \left[k+1 + \sum_{i=1}^na_i , k+\sum_{i=1}^{n+1}a_i \right ] . \\ \end{cases}\ ] ] it is non - decreasing and such that for every @xmath611 we have @xmath271 . so @xmath45 is a kneading map . if @xmath612 is the corresponding sequence of transition matrices , then it is easy to see that for each @xmath90 we have @xmath613 by considering a bratteli diagram isomorphic to @xmath418 , that only differs with it in the first @xmath614 levels , we obtain @xmath615 so the dimension group associated to @xmath316 is isomorphic to direct limit @xmath616 see for example ( * ? ? ? * theorem 3.7 ) , which by ( * ? ? ? * theorem 4.8 ) is isomorphic to @xmath617 . the purpose of this appendix is to prove corollary [ c : ergodic theory holomorphic ] . as the parameters given by ( the proof of ) the main theorem are such that the kneading map of corresponding logistic map diverges to @xmath49 , this result is a direct consequence of the following lemma . [ l : ergodic theory holomorphic ] let @xmath17 $ ] be a parameter such that the kneading map of @xmath12 diverges to @xmath48 . consider the corresponding quadratic polynomial @xmath28 , and denote by @xmath31 the hausdorff dimension of the julia set of @xmath28 . then , for an invariant probability measure of @xmath28 that is supported on the julia set of @xmath28 , the following properties are equivalent . as the restriction of @xmath28 to @xmath10 $ ] is the logistic map @xmath12 , the implication @xmath618 is given by ( * ? ? ? * lemma 21 ) . the proof of the implication @xmath619 is analogous to the corresponding implication of the same lemma . the implication @xmath620 being trivial we just need to prove the implication @xmath621 to complete the proof of the lemma . we will show that for each invariant measure @xmath21 that is supported on the julia set of @xmath28 , we have @xmath622 with equality when the lyapunov exponent of @xmath21 is zero . for such a measure we have @xmath623 by @xcite , so there are two cases . if @xmath624 , then @xmath21 is supported on the post - critical set of @xmath28 and therefore we have @xmath625 by the variational principle and the fact that the topological entropy of @xmath28 restricted to its post - critical set is zero ( proposition [ p : continuous unimodal ] ) . so we have @xmath626 in this case . suppose now that @xmath627 . then by @xcite it follows that , if we denote by @xmath628 the hausdorff dimension of @xmath21 , then @xmath629 . as @xmath630 , we obtain @xmath622 in this case . h. bruin . combinatorics of the kneading map . in _ thirty years after sharkovski s theorem : new perspectives ( murcia , 1994 ) _ , volume 8 of _ world sci . nonlinear sci . theme issues proc . _ , pages 7787 . world sci . publ . , river edge , nj , 1995 . reprint of the paper reviewed in mr1361922 ( 96k:58070 ) . ricardo ma . the hausdorff dimension of invariant probabilities of rational maps . in _ dynamical systems , valparaiso 1986 _ , volume 1331 of _ lecture notes in math . _ , pages 86117 . springer , berlin , 1988 .	a well - known consequence of the ergodic decomposition theorem is that the space of invariant probability measures of a topological dynamical system , endowed with the weak@xmath0 topology , is a non - empty metrizable choquet simplex . we show that every non - empty metrizable choquet simplex arises as the space of invariant probability measures on the post - critical set of a logistic map . here , the post - critical set of a logistic map is the @xmath1-limit set of its unique critical point . in fact we show the logistic map @xmath2 can be taken in such a way that its post - critical set is a cantor set where @xmath2 is minimal , and such that each invariant probability measure on this set has zero lyapunov exponent , and is an equilibrium state for the potential @xmath3 . logistic map , post - critical set , invariant measures , choquet simplices , minimal cantor system , generalized odometer 37e05 , 37a99 , 37b10 , 54h20
for @xmath0 , a self - gravitating system is expected to achieve a statistical equilibrium state described by the boltzmann distribution . this statistical approach is the correct description of globular clusters , which are small groups of stars ( @xmath3 ) whose age is of the same order as the relaxation time . their observed thermal distribution ( michie - king models ) is the natural outcome of a `` collisional '' relaxation due to the development of stellar encounters . this collisional relaxation is usually described by the gravitational kinetic landau equation which satisfies a h - theorem for the boltzmann entropy . fundamentally , the boltzmann entropy is equal to the logarithm of disorder where the disorder measures the number of microstates ( complexions ) associated with a given macrostate . the maximization of the boltzmann entropy at fixed mass and energy has a clear statistical and thermodynamical meaning : it determines the most probable distribution of stars at equilibrium , assuming that the microstates are equiprobable . however , unlike ordinary systems , the statistical mechanics of self - gravitating systems encounters some difficulties due to the long - range unshielded nature of gravity and the divergence of the potential at short distances @xcite . in particular , there is no statistical equilibrium state in a strict sense : the boltzmann entropy has no global maximum and isothermal distributions have infinite mass @xcite . the increase of entropy as the system spreads reflects the natural tendency of stellar clusters to evaporate ( this is already the case for a classical gas , without gravity , if it is not confined by the walls of a container ) . a statistical equilibrium state can be defined only in a weaker sense if the system is tidally truncated or confined by the pressure exerted by an external medium . theorists overcome the infinite mass problem by confining the system within a box . physically , the box delimits the region of space where the system can be assumed isolated and described by thermodynamical arguments . typically , the box radius plays the same role as the tidal radius . however , even for box - confined systems , statistical equilibrium does not exist if the energy is too low . below the antonov threshold @xcite , the system undergoes gravitational collapse called gravothermal catastrophe @xcite . this core collapse ultimately leads to the formation of binaries @xcite . statistical equilibrium states exist at sufficiently high energies . they are metastable ( local entropy maxima ) but their lifetime is considerable as it increases exponentially rapidly with the number of particles @xcite . therefore , small clusters of stars such as globular clusters correspond to long - lived metastable states . furthermore , for self - gravitating systems , statistical ensembles are non - equivalent @xcite and , for isolated systems , only the microcanonical ensemble makes sense . to our opinion , the evaporation of stellar clusters , the gravothermal catastrophe , the formation of binaries and the long - lived metastability of globular clusters are the correct answers to the absence of strict statistical equilibrium for self - gravitating systems . they correspond to important physical processes that have been clearly identified in astrophysics . _ with these limitations in mind , self - gravitating systems can be described by conventional statistical mechanics @xcite . _ however , for systems with long - range interactions , the collisional relaxation time is extremely long because it increases algebraically with the number @xmath1 of particles . for example , for stellar systems , the chandrasekhar relaxation time scales as @xmath4 , where @xmath5 is the dynamical time . therefore , for a very long period , encounters between stars are negligible and the galactic dynamics is described by the vlasov , or collisionless boltzmann , equation coupled to gravity via the poisson equation @xcite . the vlasov - poisson system is the correct description of large groups of stars such as elliptical galaxies ( @xmath6 ) that are still , at the present time , in the collisionless regime ( contrary to globular clusters ) . now , the vlasov equation admits an infinite class of stationary solutions . along its collisionless evolution , a stellar system can be trapped in one of these stationary solutions and remain frozen in that state for a very long time , until collisional effects come into play . physically , the convergence to a steady state is due to phase mixing and incomplete violent relaxation @xcite . these steady states , referred to as quasi - stationary states or metaequilibrium states , are described by non - standard ( non - boltzmannian ) distributions @xmath7 . this is not really surprising since the system is collisionless so that there is no relation with thermodynamics in the usual sense . tsallis distributions @xmath8 @xcite form a particular class of stationary solutions of the vlasov equation . in fact , these distributions are known for a long time in astrophysics . they are called _ stellar polytropes _ and were introduced by plummer @xcite in 1911 and eddington @xcite in 1916 . although these distributions have interesting mathematical properties and have been used to construct simple theoretical models of galaxies @xcite , it is important to note that they do not provide a good fit of observed galaxies . once again , this is not really surprising : there is no fundamental reason why polytropic distributions ( or tsallis distributions ) should be selected among other stationary solutions of the vlasov equation @xcite . any distribution function @xmath9 , where @xmath10 is the stellar energy , is a stationary solution of the vlasov equation . it extremizes a functional of the form @xmath11=-\int c(f)d^{3}{\bf r}d^{3}{\bf v}$ ] , where @xmath12 is convex , at fixed mass @xmath13 and energy @xmath14 . such functionals have been called h - functions by tremaine , hnon & lynden - bell @xcite . tsallis functional @xmath15 $ ] is the h - function associated with stellar polytropes @xcite . furthermore , the maximization of a h - function at fixed mass and energy provides a criterion of nonlinear dynamical stability for the vlasov equation @xcite . this `` dynamical interpretation '' defended by @xcite and @xcite who view metaequilibrium states as stationary solutions of the vlasov equation and interpret the maximization of a @xmath16-function as a condition of nonlinear dynamical stability is different from the `` generalized thermodynamical interpretation '' defended by @xcite who view tsallis functional @xmath15 $ ] as a generalized entropy , interpret quasi - stationary states as generalized maximum entropy states and consider the maximization of @xmath15 $ ] as a condition of generalized thermodynamical stability.=-{1\over q-1}\int ( f^{q}-f)d{\bf r}d{\bf v}$ ] defined in terms of the distribution function @xmath17 and tsallis entropy @xmath18=-{1\over q-1}\int ( \rho^{q}-\rho)d\eta d{\bf r}d{\bf v}$ ] defined in terms of the probability density @xmath19 of phase levels . ignoring the specificities of the collisionless evolution ( casimir constraints ) , the authors of @xcite interprete @xmath15 $ ] as a generalized entropy while the correct form of tsallis entropy in the context of violent relaxation is @xmath18 $ ] ( it returns the lynden - bell entropy for @xmath20 which is the appropriate form of boltzmann entropy in that context ) . by contrast , @xcite interprete the tsallis functional @xmath15 $ ] as a @xmath16-function . in this paper , we only consider the functional @xmath15 $ ] . the important distinction between the functionals @xmath15 $ ] and @xmath18 $ ] is further discussed in @xcite . ] there are two points that we would like to emphasize in this paper . the first point is that tsallis ( and boltzmann ) distributions do not have a fundamental importance in regard with the collisionless dynamics of stellar systems ( when the @xmath21 limit is taken before the @xmath0 limit ) . these are just _ particular _ stationary solutions of the vlasov equation @xcite . this is different from the boltzmann distribution in regard with the collisional dynamics of stellar systems ( when the @xmath0 limit is taken before the @xmath22 limit ) . the boltzmann distribution is of fundamental importance as it describes the statistical equilibrium state of the system ( up to all the limitations mentioned in the first paragraph ) . furthermore , it is the _ unique _ stationary solution of the kinetic landau equation which describes the collisional relaxation of stellar systems to order @xmath23 @xcite . thus , the first order correction to the vlasov limit selects the boltzmann distribution . the second point that we want to emphasize is that , formally , the criterion of nonlinear dynamical stability for collisionless stellar systems described by the vlasov - poisson system _ resembles _ a criterion of generalized thermodynamical stability in the microcanonical ensemble . therefore , we can develop an _ effective thermodynamical formalism _ ( e.t.f ) to investigate the nonlinear dynamical stability of collisionless stellar systems @xcite . in this _ thermodynamical analogy _ , the notions of entropy , free energy and temperature are effective . we stress that our e.t.f . has a meaning different from the one given by @xcite . in our point of view , tsallis functional @xmath15 $ ] is related to dynamics , not thermodynamics , in the case of collisionless stellar systems . what we do essentially is to use the language of thermodynamics to study nonlinear dynamical stability problems . we stress that the `` thermodynamical analogy '' that we develop is only valid for distribution functions of the form @xmath24 with @xmath25 that describe a sub - class of _ spherical _ stellar systems @xcite . we have also found that , formally , the criterion of nonlinear dynamical stability for barotropic stars described by the euler - poisson system resembles a criterion of thermodynamical stability in the canonical ensemble @xcite . since `` canonical stability implies microcanonical stability in thermodynamics '' , we deduce , using our thermodynamical analogy , that `` the stability of a barotropic star implies the stability of the corresponding stellar system ( but not the converse ) '' . this provides an original interpretation of the nonlinear antonov first law in terms of ensemble inequivalence @xcite . tsallis distributions can also arise in other domains of physics and astrophysics for different reasons . for example , polytropic distributions correspond to the unique stationary state of a certain class of generalized kinetic equations or generalized fokker - planck equations satisfying a h - theorem for the tsallis entropy ( or free energy ) @xcite . however , it is not clear at the present time to which physical systems these generalized kinetic equations apply . we have argued that they may be `` effective equations '' trying to take into account `` hidden constraints '' in complex systems @xcite . we have also shown that it is possible to formally generalize kinetic and fokker - planck equations , so that they increase an arbitrary entropy @xcite , not necessarily the boltzmann or the tsallis one . in a different context , taruya & sakagami @xcite have shown that the collisional evolution of stellar systems , between the phase of violent relaxation and the phase of gravothermal catastrophe , can be described by a sequence of polytropic distributions with a time varying index , interpreted as _ quasi - equilibrium states_. in conclusion , tsallis distributions ( and other non - standard distributions ) arise in different domains of physics ( biology , economy , ... ) for reasons that are not necessarily related to thermodynamics . the important point is to explain , in each specific situation , what is the _ physical meaning _ of such distributions . we shall give here various examples in astrophysics but the concepts that we discuss can apply to other systems with long - range interactions . we consider a self - gravitating gaseous medium described by the euler - poisson system @xmath26 @xmath27 @xmath28 we assume that the gas is barotropic with an equation of state @xmath29 . this provides the simplest model of stars @xcite . the total energy of the star is @xmath30=\int\rho\int_{0}^{\rho}{p(\rho')\over\rho^{'2}}\ , d\rho'd^{3}{\bf r}+{1\over 2}\int \rho\phi\ , d^{3}{\bf r}+\int \rho { { \bf u}^{2}\over 2}\,d^{3}{\bf r}. \label{ep4}\ ] ] the first term is the work @xmath31 done in compressing it from infinite dilution , the second term is the gravitational energy and the third term is the kinetic energy associated with the mean motion . it is straightforward to verify that the energy functional ( [ ep4 ] ) is conserved by the euler - poisson system ( @xmath32 ) . therefore , a minimum of @xmath33 at fixed mass determines a stationary solution of the euler - poisson system which is nonlinearly dynamically stable @xcite . physically , this means that a small perturbation remains close ( in some norm ) to the minimum . we are led therefore to consider the minimization problem @xmath34\quad \|\ m[\rho]=m\rbrace . \label{ej4}\end{aligned}\ ] ] this minimization problem is consistent with the phenomenology of star formation . a star forms through the gravitational collapse of a molecular cloud until it has reached a state of minimum energy . canceling the first order variations of eq . ( [ ep4 ] ) , we obtain @xmath35 and the condition of hydrostatic equilibrium @xmath36 between pressure and gravity . therefore , extrema of @xmath33 correspond to stationary solutions of the euler - poisson system ( [ ep1])-([ep3 ] ) . combining the condition of hydrostatic equilibrium ( [ ep5 ] ) and the equation of state @xmath29 , we get @xmath37 so that @xmath38 is a function of @xmath39 that we note @xmath40 . considering now the second order variations , the condition of nonlinear dynamical stability is @xmath41 for all perturbations that conserve mass , i.e. @xmath42 . we note that the second integral in eq . ( [ ep7 ] ) can be written in a more conventional form by using @xmath43 . the minimization problem ( [ ej4 ] ) has been studied in detail in @xcite for spherically symmetric systems described by an isothermal and a polytropic equation of state ( the system can be self - confined or confined by a box or by an external pressure ) . it is shown that the point in the series of equilibria where the system ceases to be a minimum of @xmath33 and becomes a saddle point coincides with the point of neutral linear stability . therefore , the conditions of linear and nonlinear dynamical stability coincide . in the theory of stellar structure , it is usually assumed ( at least in the simple models that we consider here ) that a star is a perfect gas in local thermodynamic equilibrium ( l.t.e ) . therefore , at each point , the distribution function of the particles is of the form @xmath44 if we define the pressure by @xmath45 where @xmath46 , we find that the local equation of state is @xmath47 we shall now consider two types of models that have been widely studied in astrophysics , namely the case of isothermal stars and the case of polytropic stars . _ ( i ) isothermal stars : _ if we assume that the star is in thermal equilibrium , then the temperature is uniform : @xmath48 . from one place to the other the transformation is _ isothermal_. in that case , the equation of state becomes @xmath49 since the pressure depends only on the density , the fluid is barotropic . the energy functional ( [ ep4 ] ) reads @xmath50=k_{b}t\int { \rho\over m}\ln { \rho\over m}\ , d^{3}{\bf r } + { 1\over 2}\int \rho\phi\ , d^{3}{\bf r}+\int \rho { { \bf u}^{2}\over 2}\,d^{3}{\bf r}. \label{ip1}\ ] ] we note that the energy functional ( [ ip1 ] ) of an isothermal gas coincides with the boltzmann free energy @xmath51 of a self - gravitating system in the canonical ensemble ( see @xcite for more details ) . furthermore , the distribution function of the particles is the maxwell - boltzmann distribution ( with a uniform temperature ) . the relation ( [ ep6 ] ) between the density and the potential can be explicitly written @xmath52 according to eq . ( [ ej4 ] ) , it can be obtained by extremizing eq . ( [ ip1 ] ) at fixed mass or by combining the condition of hydrostatic equilibrium with the isothermal equation of state . _ ( ii ) polytropic stars : _ stars are not isothermal in general . in the convective region , they are rather in a situation where the entropy is uniform : @xmath53 . from one place to the other the transformation is _ isentropic_. in that case , using the first principle of thermodynamics @xmath54 with @xmath55 and @xmath56 , we obtain the differential equation @xmath57 ( with @xmath58 ) . using the equation of state of a perfect gas @xmath59 and the mayer relation @xmath60 , the differential equation can be integrated and yields the polytropic equation of state @xmath61 where @xmath62 is the polytropic constant and @xmath63 is the ratio of specific heats at constant pressure and constant volume @xcite . the polytropic index @xmath64 is defined by @xmath65 . the polytropic constant can be written @xmath66 , where @xmath67 is sometimes called a polytropic temperature ( cf . @xcite , p. 86 ) . for @xmath68 , we recover the isothermal case with @xmath69 and @xmath70 . for a polytropic gas , the local temperature @xmath71 , defined by ( [ lte2 ] ) , is given by @xmath72 we note that the temperature is position dependent ( while the specific entropy @xmath73 is uniform ) and related to the density by a power law ; this is the local version of the usual isentropic law @xmath74 in standard thermodynamics . the polytropic index @xmath64 is related to the gradients of temperature and density according to @xmath75 the energy functional ( [ ep4 ] ) can be written @xmath50={k\over \gamma-1}\int ( \rho^{\gamma}-\rho)\,d^{3}{\bf r } + { 1\over 2}\int \rho\phi \,d^{3}{\bf r}+\int \rho { { \bf u}^{2}\over 2}\,d^{3}{\bf r}. \label{ip4}\ ] ] for an isentropic transformation ( @xmath55 ) , the first term in eqs . ( [ ep4 ] ) and ( [ ip4 ] ) represents the internal energy since @xmath76 . for the polytropic equation of state ( [ ip0b ] ) , the internal energy is a power - law functional : @xmath77 . we have then added a constant term @xmath78 , proportional to the total mass , in the polytropic energy functional ( [ ip4 ] ) so as to recover the isothermal energy functional ( [ ip1 ] ) for @xmath79 ( i.e. , @xmath68 ) . using eq . ( [ ep6 ] ) or extremizing the energy functional ( [ ip4 ] ) at fixed mass , we find that the relation between the density and the potential is @xmath80 we see at this stage some resemblances with tsallis distributions and free energies , see eqs . ( [ ip5 ] ) and ( [ ip4 ] ) . however , there is _ no _ generalized thermodynamics in the problem . the physics of gaseous stars can be understood in terms of conventional thermodynamics once realized that adiabatic transformations are more relevant than isothermal transformations in stellar structure . we shall see that these formal resemblances are essentially fortuitous or the mark of a _ for @xmath0 , a self - gravitating system is expected to achieve a statistical equilibrium state . assuming that the accessible microstates are equiprobable , the most probable macroscopic distribution function maximizes the boltzmann entropy @xmath81=-\int { f\over m}\ln { f\over m}\,d^{3}{\bf r}d^{3}{\bf v } , \label{coll1}\ ] ] at fixed mass and energy . the boltzmann entropy can be obtained by a standard combinatorial analysis . the criterion of thermodynamical stability is therefore @xmath82\quad \|\ e[f]=e , m[f]=m \rbrace.\ ] ] it determines the most probable distribution of stars at statistical equilibrium . introducing lagrange multipliers and writing @xmath83 , we get the boltzmann distribution @xmath84 we note that the statistical equilibrium state depends only on the energy @xmath85 of the stars ( for a non - rotating system ) . the collisional evolution is usually described by the kinetic landau equation , see e.g. @xcite , that we write symbolically as @xmath86 where @xmath87 is the self - consistent gravitational field produced by the stars and @xmath88 is the landau collision term . the landau - poisson system conserves mass and energy and increases the boltzmann entropy ( h - theorem ) . the boltzmann distribution is the only stationary solution of this equation . from the kinetic theory of stellar systems , we can show that the collisional relaxation time ( chandrasekhar s time ) scales as @xmath89 . for @xmath90 , we can ignore the collision term in the kinetic equation . in this regime , a stellar system is described by the vlasov - poisson system @xmath91 @xmath92 the vlasov description assumes that the evolution of the system is encounterless . this is a very good approximation for the dynamics of stars in elliptical galaxies ( @xmath6 ) . for smaller groups of stars , such as globular clusters ( @xmath3 ) , encounters have to be taken into account and the dynamics is governed by the landau - poisson system or by the orbit averaged fokker - planck equation @xcite . here , we exclusively focus on the collisionless dynamics . it is thus clear from the beginning that thermodynamics , in its usual sense , is irrelevant to the present context ; it makes sense only on much longer timescales when stellar encounters have to be taken into account . the vlasov equation admits an infinite number of stationary solutions @xcite . during its collisionless evolution , a stellar system can be trapped in a stable stationary state and remains frozen in that quasi - stationary state for a very long time ( until collisions come into play and drive the slow relaxation ) . according to the jeans theorem , the general form of stationary solutions of the vlasov equation is a function @xmath93 of the integrals of motion . of course , stable _ stationary solutions are of physical interest . in the general case , it is difficult to derive a criterion of dynamical stability . we shall restrict ourselves in this paper to stationary distribution functions of the form @xmath24 which depend only on the energy of the stars @xmath94 . such distribution functions describe a sub - class of _ spherical stellar systems _ ( in the general case , @xmath95 depends on energy @xmath96 and angular momentum @xmath97 ) @xcite . for such distributions , it is possible to derive a general criterion of nonlinear dynamical stability . we consider the class of functionals @xmath98=-\int c(f)\,d^{3}{\bf r}d^{3}{\bf v},\label{vh2}\ ] ] where @xmath12 is an arbitrary convex function . these functionals are called h - functions @xcite because they share some analogies with the boltzmann h - function in kinetic theory ( see the conclusion).-functions are defined with the coarse - grained distribution function @xmath99 . we shall argue in the conclusion and in @xcite that , in case of violent relaxation , it is the _ coarse - grained _ distribution that is a nonlinearly dynamically stable stationary solution of the vlasov equation . therefore , in all the paper , @xmath95 must be regarded in fact as the coarse - grained distribution function @xmath100 . naturally , the stability results also apply to a steady distribution function @xmath95 ( fine - grained ) , but this supposes that the system has been initially prepared in such a stationary solution of the vlasov equation , which is relatively artificial . ] since these functionals are particular casimirs , they are conserved by the vlasov equation . the total energy @xmath101 and the total mass @xmath102=\int fd^{3}{\bf r}d^{3}{\bf v}$ ] are also conserved by the vlasov equation . therefore , a maximum of @xmath103 at fixed mass @xmath13 and energy @xmath14 determines a stationary solution @xmath17 of the vlasov equation that is nonlinearly dynamically stable . we are led therefore to consider the maximization problem @xmath104\quad \|\ e[f]=e , m[f]=m \rbrace.\ ] ] we note that @xmath105=e[f]-ts[f]$ ] , where @xmath106 is an arbitrary positive constant , is also conserved by the vlasov equation ( this is called an energy - casimir functional @xcite ) . therefore , a minimum of @xmath107 at fixed mass @xmath13 is also a nonlinearly dynamically stable stationary solution of the vlasov equation . this criterion can be written @xmath108\quad \|\ m[f]=m \rbrace,\ ] ] where @xmath109={1\over 2}\int f v^{2}\,d^{3}{\bf r}d^{3}{\bf v}+{1\over 2}\int \rho\phi \,d^{3}{\bf r } + t\int c(f)d^{3}{\bf r}\,d^{3}{\bf v}.\ ] ] similar criteria of nonlinear dynamical stability have been introduced in the context of two - dimensional hydrodynamics described by the 2d euler - poisson system @xcite . the optimization problem ( [ vh4 ] ) corresponds to the nonlinear dynamical stability criterion of holm _ et al . _ @xcite and the optimization problem ( [ vh3 ] ) corresponds to the refined dynamical stability criterion of ellis _ et al . these criteria are in general not equivalent for systems with long - range interactions , and this is similar to a notion of ensemble inequivalence in thermodynamics ( see sec . [ sec_ta ] ) . inspired by these studies , we have introduced the same criteria to study the nonlinear dynamical stability of stellar systems described by the vlasov - poisson system @xcite . recently , we found that similar optimization principles have also been introduced independently by rein and guo in the mathematical literature , see @xcite . introducing lagrange multipliers and writing @xmath110 or @xmath111 , we find that the critical points of the variational problems ( [ vh3 ] ) and ( [ vh4 ] ) are both given by @xmath112 since @xmath113 is a monotonically increasing function of @xmath95 , we can inverse this relation to obtain @xmath114 where @xmath115 . therefore , the optimization problems ( [ vh3 ] ) and ( [ vh4 ] ) determine stationary solutions of the vlasov equation of the form @xmath24 . as said previously , this is a particular case of the jeans theorem describing a sub - class of spherical stellar systems @xcite . from the identity @xmath116 resulting from eq . ( [ vh6 ] ) , @xmath117 is a monotonic function . assuming that @xmath95 is decreasing , which is the physical situation , imposes @xmath118 . setting @xmath119 , the condition of nonlinear dynamical stability provided by the criterion ( [ vh3 ] ) is @xmath120 and the condition of nonlinear dynamical stability provided by the criterion ( [ vh4 ] ) is @xmath121 the first integral can be written in a more conventional form by using @xmath116 . we note the importance of the first order constraints in the criteria ( [ vh7 ] ) and ( [ vh8 ] ) . if condition ( [ vh8 ] ) is satisfied for all perturbations that conserve mass , it is a fortiori satisfied for perturbations that conserve mass _ and _ energy . therefore , condition ( [ vh8 ] ) implies condition ( [ vh7 ] ) , but not the opposite . we conclude that the stability criterion ( [ vh3 ] ) is more refined than the stability criterion ( [ vh4 ] ) : if a collisionless stellar system satisfies ( [ vh3 ] ) or ( [ vh4 ] ) , it is nonlinearly dynamically stable ; however , if it does not satisfy ( [ vh4 ] ) , it can be nonlinearly dynamically stable provided that it satisfies ( [ vh3 ] ) . this means that we can `` miss '' stable solutions if we use just the optimization problem ( [ vh4 ] ) . the problem ( [ vh3 ] ) is richer and allows to construct a larger class of nonlinearly dynamically stable models . in particular , complete stellar polytropes with index @xmath122 satisfy the criterion ( [ vh3 ] ) but not the criterion ( [ vh4 ] ) @xcite . to study the nonlinear dynamical stability of a collisionless stellar system described by a distribution function of the form @xmath24 with @xmath25 , we are led to consider the optimization problems ( [ vh3 ] ) and ( [ vh4 ] ) . we note that , formally , they are _ similar _ to criteria of thermodynamical stability for collisional self - gravitating systems , see eq . ( [ coll2 ] ) , but they involve a more general functional @xmath11=-\int c(f)\,d^{3}{\bf r}d^{3}{\bf v}$ ] ( h - function ) than the boltzmann entropy ( [ coll1 ] ) . furthermore , they have a completely different interpretation as they provide criteria of nonlinear dynamical stability for a stationary solution of the vlasov equation ( representing a quasi - stationary structure in the collisionless regime ) while the maximization of the boltzmann entropy at fixed mass and energy provides a condition of thermodynamical stability for the statistical equilibrium state reached for @xmath0 ( representing the most probable state that we may expect ) . due to the formal resemblance between the nonlinear dynamical stability criteria and the thermodynamical stability criteria , we can develop a _ thermodynamical analogy _ @xcite to investigate the nonlinear dynamical stability of collisionless stellar systems with @xmath24 and @xmath25 . in this analogy , @xmath103 plays the role of an `` effective entropy '' , @xmath123 plays the role of an `` effective temperature '' and @xmath107 plays the role of an `` effective free energy '' ( it is related to @xmath103 by a legendre transform ) . the criterion ( [ vh3 ] ) is similar to a condition of `` microcanonical stability '' and the criterion ( [ vh4 ] ) is similar to a condition of `` canonical stability '' . since canonical stability implies microcanonical stability ( but not the converse ) , we recover the fact that condition ( [ vh8 ] ) implies condition ( [ vh7 ] ) . the stability problems ( [ vh3 ] ) and ( [ vh4 ] ) can be studied by plotting the linear series of equilibria @xmath124 , which is similar to a caloric curve in thermodynamics . according to poincar s theorem , a mode of stability is lost or gained at a turning point or at a branching point . linear series of equilibria have been studied in detail in relation with the thermodynamics of self - gravitating systems for the boltzmann @xcite and fermi - dirac @xcite entropies . due to the above thermodynamical analogy , the method of linear series of equilibria can be used similarly to settle the nonlinear dynamical stability of a collisionless stellar system for a general @xmath16-function . note that the slope @xmath125 ( where @xmath126 is the lagrange multiplier associated with the energy ) plays an important role because , according to the turning point argument , its change of sign is the signal of an instability . in thermodynamics , @xmath127 represents the specific heat . isothermal self - gravitating spheres cease to be minima of free energy when @xmath125 becomes negative passing by @xmath128 , and they cease to be maxima of entropy when @xmath125 becomes positive again passing by @xmath129 @xcite@xcite . therefore , the criteria of canonical and microcanonical stability are inequivalent in the region of negative specific heats @xmath130 . due to the thermodynamical analogy , we have similar results for the nonlinear dynamical stability of collisionless stellar systems ( for an arbitrary h - function ) but with a different interpretation . for example , the series of equilibria for polytropic distributions has been studied in @xcite . in this context , the sign of the slope @xmath125 , which is similar to an effective specific heat , is connected to the nonlinear dynamical stability of the system ( see fig . [ el4 ] ) . _ this `` dynamical interpretation '' @xcite is different from the `` generalized thermodynamical interpretation '' of taruya & sakagami @xcite who considered the same mathematical problem in relation with tsallis generalized thermodynamics . _ for any spherical stellar system with @xmath24 , there exists a corresponding barotropic star with the same equilibrium density distribution . indeed , defining the density and the pressure by @xmath131 , @xmath132 , and eliminating the potential @xmath39 between these two expressions , we find that @xmath29 . writing explicitly the density and the pressure in the form @xmath133 and @xmath134 , and taking the derivative of the second equation , we obtain the condition of hydrostatic equilibrium ( [ ep5 ] ) . as discussed in sec . [ sec_vh ] , the minimization principle ( [ vh4 ] ) provides a _ sufficient _ condition of nonlinear dynamical stability for a collisionless stellar system . to solve this minimization principle , we can first minimize @xmath105 $ ] at fixed density @xmath135 to obtain @xmath136 . we can then express the functional @xmath105 $ ] as a functional of the density @xmath135 by setting @xmath137=f[f_{*}]$ ] . the calculations are detailed in @xcite and the functional @xmath137 $ ] can be finally written @xmath138={1\over 2}\int\rho\phi\ , d^{3}{\bf r}+\int\rho\int_{0}^{\rho}{p(\rho')\over \rho'^{2}}\,d\rho'd^{3}{\bf r } , \label{anto4}\end{aligned}\ ] ] where @xmath139 is the equation of state of the corresponding barotropic gas . we are now led to consider the minimization problem @xmath140\quad \|\ m[\rho]=m \rbrace . \label{anto3}\end{aligned}\ ] ] now , we observe that the functional ( [ anto4 ] ) coincides with the energy functional ( [ ep4 ] ) of a barotropic gas with @xmath141 . we arrive therefore at the following conclusion : for a given stellar system with @xmath24 and @xmath25 , if we know that the corresponding barotropic star is nonlinearly dynamically stable with respect to the euler - poisson system , then it is a minimum of @xmath142 $ ] , hence of @xmath137 $ ] ( at fixed mass ) . therefore , the stellar system satisfies the criterion ( [ vh4 ] ) so it is nonlinearly dynamically stable with respect to the vlasov - poisson system . this leads to a nonlinear generalization of the antonov first law : `` a stellar system is nonlinearly dynamically stable whenever the corresponding barotropic gas is nonlinearly dynamically stable '' @xcite . however , the reciprocal is wrong in general . a stellar system can be nonlinearly dynamically stable according to the `` microcanonical '' criterion ( [ vh3 ] ) while it does not satisfy the `` canonical '' criterion ( [ vh4 ] ) so that the corresponding barotropic star is dynamically unstable @xcite . due to the thermodynamical analogy , the antonov first law for collisionless stellar systems has the same status as the fact that : `` canonical stability implies microcanonical stability in thermodynamics '' . the reciprocal is wrong if the ensembles are not equivalent . to the point of view of their dynamical stability , the crucial difference between stars and galaxies is that , for galaxies , both @xmath103 ( h - function ) and @xmath14 are _ individually _ conserved by the vlasov equation while in the case of barotropic stars only the total energy @xmath33 is conserved by the euler equations . therefore , the most refined stability criterion for galaxies is ( [ vh3 ] ) while the most refined stability criterion for stars is ( [ ej4 ] ) , which is included in ( [ vh3 ] ) . this is the intrinsic reason why a spherical galaxy can be stable even if the corresponding barotropic star is unstable . it is interesting to consider the application of these results to polytropes ( see fig . [ el4 ] ) . by plotting the series of equilibria @xmath124 , it can be shown that a complete polytrope ( i.e. , for which the density drops to zero at a finite radius ) ceases to be a minimum of @xmath107 for @xmath143 @xcite . this is when the slope @xmath125 becomes negative in the series of equilibria . in the thermodynamical analogy , this is similar to a loss of canonical stability . using the above result , we conclude that polytropic stars are nonlinearly dynamically stable for @xmath144 and they become dynamically unstable for @xmath143 @xcite . however , complete polytropes cease to be a maximum of @xmath103 only for @xmath145 ( they also cease to be self - confined at this index ) @xcite . this is similar to a loss of microcanonical stability in the thermodynamical analogy . therefore , stellar polytropes are nonlinearly dynamically stable for @xmath145 and they become dynamically unstable for @xmath146 . for @xmath147 , stellar polytropes are nonlinearly dynamically stable while corresponding polytropic stars are dynamically unstable . this is similar to a situation of ensemble inequivalence ( in a region of negative specific heats ) in thermodynamics @xcite . of course , the dynamical stability of stellar polytropes and polytropic stars has been established for a long time in astrophysics @xcite but the interpretation of these results ( and of the antonov first law ) in terms of a thermodynamical analogy and a situation of ensemble inequivalence is new . furthermore , only linear dynamical stability is considered in @xcite while the criteria ( [ vh3 ] ) and ( [ vh4 ] ) are criteria of _ nonlinear _ dynamical stability . similar criteria of nonlinear dynamical stability can be developed for the hmf model @xcite ( in that case , the `` ensembles '' are equivalent ) and for other systems described by vlasov or euler equations @xcite . we consider the @xmath16-function @xmath148 this functional resembles the boltzmann entropy ( [ coll1 ] ) in thermodynamics . however , as explained above , its physical interpretation is different . its maximization at fixed mass and energy determines a nonlinearly dynamically stable stationary solution of the vlasov equation with distribution function @xmath149 this distribution function has the same form ( but a different interpretation ) as the statistical equilibrium state ( [ coll3 ] ) of a stellar system resulting from a collisional relaxation or as the distribution function ( [ lte1 ] ) of the gas in an isothermal star . for these reasons , collisionless stellar systems described by eq . ( [ vb2 ] ) are sometimes called isothermal stellar systems and @xmath150 is sometimes called a `` temperature '' . in fact , it is more proper to regard @xmath106 as the velocity dispersion of the stellar system . we note also that the mass @xmath151 of the individual particles does not appear in the h - function ( [ vb1 ] ) and in the distribution function ( [ vb2 ] ) contrary to the case of statistical equilibrium , see ( [ coll1 ] ) and ( [ coll3 ] ) . this is because the present description is based on the vlasov equation in which the mass of the particles does not appear . therefore , the distribution function ( [ vb2 ] ) does not lead to a segregation by mass contrary to the statistical equilibrium distribution of a multi - species self - gravitating system @xcite . the barotropic gas associated with the stellar system defined by eq . ( [ vb2 ] ) is the isothermal gas with an equation of state @xmath152 where @xmath123 . the functional ( [ anto4 ] ) takes the form @xmath138={1\over 2}\int \rho\phi\ , d^{3}{\bf r}+t\int \rho\ln\rho \,d^{3}{\bf r}. \label{vb4}\end{aligned}\ ] ] it coincides with the energy functional ( [ ip1 ] ) of an isothermal star with @xmath35 . the density is related to the potential according to @xmath153 this relation can be directly obtained by extremizing @xmath137 $ ] at fixed mass . we can express the distribution function in terms of the density according to @xmath154 we consider the @xmath16-function @xmath155 it has the same form as the tsallis entropy in non - extensive thermodynamics @xcite . however , as explained above , its physical interpretation is different . in the present context , its maximization at fixed mass and energy determines a nonlinearly dynamically stable stationary solution of the vlasov equation , called a stellar polytrope , with distribution function @xmath156 the polytropic index is related to the parameter @xmath157 by @xmath158 stellar polytropes have been introduced by plummer @xcite in 1911 . in fact , plummer did not write the distribution function explicitly and analyzed the distribution of stars in globular clusters by treating ( incorrectly ) the system as a gas in convective equilibrium . plummer s law corresponds to the analytical solution of a polytropic gas with index @xmath159 found by schuster in 1883 . eddington @xcite in 1916 criticized the `` adiabatic '' model of plummer and justified plummer s law as a particular solution of the vlasov equation satisfying the jeans theorem . eddington was apparently the first to write the distribution function of a `` stellar polytrope '' correctly . note that the `` mistake '' of plummer is interesting because it reflects the mathematical property that `` for any spherical stellar system with @xmath24 there exists ( formally ) a corresponding barotropic star '' @xcite . later on , the fact that polytropic distributions ( [ vb7 ] ) maximize a certain casimir functional at fixed mass and energy , and that this maximization problem is related to their dynamical stability with respect to the vlasov equation has been recognized by several authors , e.g. ipser @xcite in 1974 . however , in the work @xcite the functional is simply written as @xmath160 . therefore , when @xmath68 , the connection with the functional ( [ vb1 ] ) describing isothermal stellar systems is not direct . the functional ( [ vb6 ] ) introduced by tsallis @xcite ( albeit in a quite different context ) is better to make this connection because , using lhpital s rule , eq . ( [ vb6 ] ) reduces to eq . ( [ vb1 ] ) for @xmath161 ( or @xmath68 ) . similarly , the distribution function written in the form ( [ vb7 ] ) reduces to the distribution function ( [ vb2 ] ) for @xmath162 ( or @xmath68 ) . the form given by eddington @xcite and by binney & tremaine @xcite is less convenient to make this connection . _ therefore , in the context of vlasov systems , we interpret tsallis functional as a useful h - function connecting continuously stellar polytropes and isothermal stellar systems . _ it is shown in @xcite that physical polytropic distribution functions have @xmath163 and @xmath164 . we need therefore to consider two cases according to the sign of @xmath165 . for @xmath166 ( @xmath167 ) , the distribution function can be written @xmath168 where we have set @xmath169 and @xmath170 . it is valid for @xmath171 . for @xmath172 , we set @xmath173 . this situation corresponds to the usual polytropes considered in astrophysics @xcite . for @xmath161 ( @xmath68 ) , we recover isothermal distributions and for @xmath174 the distribution function is a step function . this is the distribution function of a fermi gas at zero temperature describing classical white dwarf stars @xcite . note that stellar polytropes with @xmath175 exist mathematically but the distribution function @xmath117 increases with the energy so they may not be physical . the density and the pressure can be expressed as @xmath176 @xmath177 for @xmath178 , the distribution function can be written @xmath179 where we have set @xmath180 and @xmath181 . in that case , @xmath182 . we shall only consider distributions for which the density and the pressure ( first and second velocity moments ) are defined . this implies @xmath183 ( i.e. , @xmath184 ) . then , the density and the pressure can be expressed as @xmath185 @xmath186 eliminating the gravitational potential between the expressions ( [ vb9])-([vb10 ] ) and ( [ vb12])-([vb13 ] ) giving the density and the pressure , one finds that @xmath187 where @xmath188 @xmath189 therefore , a stellar polytrope has the same equation of state as a polytropic star . _ however , we stress that they do not have the same distribution function , except for @xmath68 ( isothermal system)_. the distribution function of the gas in a polytropic star is given by the local thermodynamic equilibrium ( [ lte1 ] ) with a space dependent temperature ( [ ip6 ] ) while the distribution function of a stellar polytrope is given by ( [ vb7 ] ) which is completely different . the functional ( [ anto4 ] ) can be put in the form @xmath138={1\over 2}\int \rho\phi \,d^{3}{\bf r}+{k\over\gamma-1 } \int ( \rho^{\gamma}-\rho)\ , d^{3}{\bf r}. \label{vb17}\end{aligned}\ ] ] it coincides with the energy functional ( [ ip4 ] ) of a polytropic star with @xmath35 . minimizing this functional at fixed mass , we directly obtain the relation between the density and the potential @xmath190 it resembles tsallis distribution ( in physical space ) where @xmath191 plays the role of the @xmath157-parameter and @xmath62 plays the role of the temperature @xmath192 ( compare eqs . ( [ vb7 ] ) and ( [ vb18 ] ) ) . also , the functional ( [ vb17 ] ) resembles tsallis free energy @xmath137=e[\rho]-ks_{\gamma}[\rho]$ ] in physical space . _ in this sense , tsallis distributions are `` stable laws '' since they keep a similar structure as we pass from phase space @xmath24 to physical space @xmath40 with the correspondence @xmath193 and @xmath194 . _ this is probably the only class of distributions enjoying this property . however , we again emphasize that , in the present context , the resemblance with a thermodynamical formalism is essentially fortuitous or effective . we can also express the distribution in terms of the density according to @xmath195 @xmath196 @xmath197 this is the counterpart of the isothermal distribution function ( [ vb5 ] ) . for any stellar system , we define the _ kinetic temperature _ by @xmath198 or @xmath199 . it is proportional to the local velocity dispersion of the particles . for a polytropic distribution function @xmath200 as indicated previously , for collisionless stellar systems @xmath106 must be regarded as a mean - squared velocity rather than a temperature ( the individual mass of the particles never appears in the equations ) . we shall use , however , the temperature terminology . using eq . ( [ vb21bis ] ) , the distribution function ( [ vb19 ] ) can be written @xmath201 @xmath202 @xmath203 note that for @xmath167 , the maximum velocity can be expressed in terms of the kinetic temperature according to @xmath204 using @xmath205 for @xmath206 , we recover the isothermal distribution ( [ vb5 ] ) for @xmath207 . on the other hand , from eqs . ( [ vb21bis ] ) and ( [ vb18 ] ) , we immediately get @xmath208 so that @xmath209 this relation can also be obtained directly from eqs . ( [ vb8 ] ) and ( [ vb23 ] ) . this shows that , for a stellar polytrope , the kinetic temperature is a linear function of the gravitational potential ( this is also true for a polytropic gas ) . the coefficient of proportionality @xmath210 is related to the polytropic index . there is an interesting application of this result related to a remark of eddington @xcite . first of all , for isolated stellar systems , the energy @xmath211 at which the distribution function vanishes should correspond to the escape energy @xmath212 . this implies @xmath213 in eq . ( [ vb8 ] ) which leads to the ordinary df of a polytrope ( the distribution ( [ vb8 ] ) with @xmath214 is in fact a generalization of the polytropic model with @xmath215 ) . now , @xmath213 implies that @xmath216 in eq . ( [ vb18 ] ) , see , e.g. , eq . ( [ vb9 ] ) . since @xmath38 and @xmath39 must then vanish at the same point , this implies that the density profile must go to @xmath217 ( with a finite mass ) which is only possible for the polytropic index @xmath159 . this is eddington s interpretation of plummer s law @xcite . on the other hand , for @xmath216 , we get @xmath218 . defining the local kinetic energy by @xmath219 and the local potential energy by @xmath220 , we get @xmath221 . thus , a _ local _ virial theorem @xmath222 holds for @xmath223 , i.e. @xmath159 . distribution functions satisfying this relation have been called hypervirial models of stellar systems @xcite . we shall now discuss different notions of temperature " that emerge in the case of polytropic distributions . in the case of isothermal distribution functions , all these notions coincide . however , for polytropic distribution functions they have a different character . \(i ) lagrange multiplier @xmath224 : the lagrange multiplier @xmath225 appearing in the variational principle @xmath110 associated with the optimization problem ( [ vh3 ] ) can be viewed as a generalized inverse temperature since @xmath226 . in terms of this lagrange multiplier @xmath225 , a polytropic distribution function can be written @xmath227 as indicated previously , the series of equilibria @xmath124 is similar to a caloric curve and the slope @xmath228 is similar to a specific heat . the sign of @xmath12 is connected to the nonlinear dynamical stability of the system via the turning point argument . this implies that the correct `` temperature '' to use in the definition of @xmath12 when using it to settle the nonlinear dynamical stability of a collisionless stellar system is the inverse of the lagrange multiplier @xmath225 . however , @xmath229 has generally _ not _ the dimension of a temperature . \(ii ) dimensional temperature : we can define a quantity that has the dimension of an inverse temperature by setting @xmath230 ( note that the normalization condition @xmath231 imposes that @xmath232 is an implicit function of @xmath225 ) . if we define furthermore @xmath233 , we can rewrite the polytropic distribution function in terms of @xmath234 as @xmath235 therefore , we see that the polytropic distribution satisfies a relation of the form @xmath236 with @xmath237 . not all distribution functions of the form @xmath238 , coming from the optimization problem ( [ vh3 ] ) , can be written as ( [ temp0c ] ) . ponno @xcite has shown in a different context that only polytropic distribution functions ( tsallis q - distributions ) satisfy this relation . \(iii ) polytropic temperature @xmath62 : we have indicated that the polytropic constant @xmath62 shares some analogies with a temperature and , for this reason , it is sometimes called a polytropic temperature . we note , in particular , that @xmath62 is uniform in a polytropic system , as is the temperature in an isothermal system . the distribution function of a stellar polytrope can be written in terms of @xmath62 as @xmath239 we note that , for a stellar polytrope , @xmath62 is a monotonic function of the lagrange multiplier @xmath225 expressed by eqs . ( [ vb15 ] ) and ( [ vb16 ] ) so they essentially play the same role . therefore , it is often convenient to use @xmath62 instead of @xmath225 when we plot the series of equilibria for polytropes ( this is the convention that we have adopted in @xcite and in fig . [ el4 ] ) . \(iv ) kinetic temperature : for any stellar system , it is natural to introduce a kinetic temperature @xmath71 through the relation @xmath198 . in general , the kinetic temperature is inhomogeneous . the distribution function of a stellar polytrope can be written in terms of @xmath71 as @xmath240 where @xmath241 . combined with eq . ( [ vb18 ] ) , this immediately yields the relation ( [ vb22 ] ) between the kinetic temperature and the gravitational potential . \(v ) energy excitation temperature : as indicated in @xcite , in any system with @xmath24 , one may define a local energy dependent excitation temperature by the relation @xmath242 for the isothermal distribution ( [ vb2 ] ) , @xmath243 coincides with the temperature @xmath123 . for the polytropic distribution ( [ vb7 ] ) , @xmath244 . this excitation temperature has a constant gradient @xmath245 related to tsallis @xmath157 parameter ( or equivalently to the index @xmath64 of the polytrope ) . the other parameter @xmath232 is related to the value of energy where the temperature reaches zero . relation ( [ temp4 ] ) is similar to that found by almeida @xcite in a different context . in sec . [ sec_stellar ] , using the orthodox interpretation of astrophysics @xcite , we have justified the polytropic distribution function ( [ vb7 ] ) as a particular stationary solution of the vlasov equation . this is probably the most relevant justification of this distribution function in astrophysics since most astrophysical bodies are governed by the collisionless boltzmann equation . in this context , the connection with a thermodynamical formalism is essentially effective ( thermodynamical analogy ) . the tsallis functional @xmath15 $ ] is a h - function whose maximization at fixed mass and energy provides a criterion of nonlinear dynamical stability . furthermore , the polytropic distribution does not play any special role among other stationary solutions of the vlasov equation . its main interest is its mathematical simplicity since it possesses properties of homology in continuity with the isothermal distribution @xcite . polytropic distribution functions can also emerge in a different context . for example , silva & alcaniz @xcite and du jiulin @xcite assume that the dynamics of their system is governed by a generalized boltzmann equation admitting a h - theorem for the tsallis entropy @xcite . in that case , the polytropic distribution is the _ only _ stationary solution of this kinetic equation . however , the relevance of this generalized kinetic equation for astrophysical systems remains to be established ( silva & alcaniz and du jiulin do not mention to which systems their approach applies ) . furthermore , as shown in chavanis @xcite , it is possible to construct even more general kinetic equations that increase a larger class of `` generalized entropies '' than the one proposed by tsallis . consequently , it has to be explained why tsallis entropy should be selected among other functionals of the form @xmath246 . concerning the theoretical study of self - gravitating polytropic spheres , our results differ from those obtained by silva & alcaniz @xcite . a first difference is that we define the kinetic temperature by @xmath247 while they use @xmath157-expectation values . another important difference is that we consider an inhomogeneous medium while silva & alcaniz consider a homogeneous system . in particular , they derive a distribution function of the form @xmath248 instead of eq . ( [ temp3 ] ) . our generalization is important because gravitational systems are inhomogeneous . thus , the equation of state associated with tsallis distribution is that of a polytrope @xmath249 while the approach of silva & alcaniz @xcite gives the impression that the equation of state is isothermal with a generalized @xmath157-temperature : @xmath250 . this is an artifact of their homogeneity assumption . in fact , for inhomogeneous self - gravitating systems , the kinetic temperature associated with a polytropic distribution is space - dependent , see eq . ( [ vb21bis ] ) . therefore , the stability condition given in @xcite based on the sign of the specific heat is not correct . the stability of polytropic spheres must take into account spatial inhomogeneity as discussed in @xcite . du jiulin @xcite considers inhomogeneous distribution functions of self - gravitating systems described by tsallis statistics . however , his approach is unnecessary complicated because he did not realize at the start that the stationary distribution function must be a function of the individual energy @xmath251 alone . indeed , for the class of generalized mean - field kinetic equations considered in @xcite , and written symbolically as @xmath252 where @xmath12 is a convex function determining the equilibrium state , a stationary solution must cancel _ independently _ the advective term and the collision term . the cancellation of the advective term implies that @xmath95 is a stationary solution of the vlasov equation . this is satisfied by any @xmath253 where @xmath107 is arbitrary . then , @xmath107 is determined by the collision term by requiring @xmath254 ( see @xcite for more details ) . we stress that the stationary solution @xmath255 of the kinetic equation ( [ gkin ] ) only depends on the energy @xmath96 while the vlasov equation admits a much larger family of stationary distributions . this can also be seen by noting that the kinetic equations in @xcite increase a generalized entropy functional @xmath256 at fixed mass and energy . therefore , the stationary state maximizes @xmath103 at fixed @xmath14 , @xmath13 and this implies @xmath257 . when @xmath103 is the tsallis entropy , this directly leads to eq . ( [ temp0 ] ) given in chavanis @xcite while du jiulin postulates the form ( [ temp3 ] ) as an extension of the homogeneous distribution ( [ silva ] ) given in silva & alcaniz @xcite . he then obtains eq . ( [ vb22 ] ) by a laborious perturbative calculation while this relation is exact and results almost immediately from eq . ( [ temp0 ] ) . in addition , du jiulin argues that this relation provides an analytic expression of the @xmath157-parameter . this is not the logical way of reading this relation . the relation ( [ vb22 ] ) is just a property of a polytropic distribution function with index @xmath157 , not an explanation of the meaning of tsallis @xmath157-parameter . equation ( [ vb22 ] ) is not mysterious ( nor fundamental ) once the physical meaning of the temperature is given . in our approach , @xmath71 represents the usual kinetic temperature related to the density and pressure via @xmath258 . said differently , @xmath259 is the local density of kinetic energy . since @xmath135 and @xmath260 are functions of the gravitational potential @xmath261 , see sec . [ sec_anto ] , this implies that @xmath71 is a function of @xmath261 . this is true for any distribution function of the form @xmath24 . now , for a polytropic distribution , @xmath262 where @xmath263 is explicitly given by eq . ( [ vb18 ] ) . this directly leads to eq . ( [ vb22 ] ) stating that @xmath106 is a linear function of @xmath39 in that case . in a recent numerical study , taruya & sakagami @xcite find that the transient stages of the collisional relaxation of the @xmath1-stars system confined within a box can be fitted by a sequence of polytropic distribution functions with a time dependent @xmath264 parameter . this is an interesting result which contrasts from the king s sequence ( truncated isothermals ) applying to tidally truncated stellar systems . therefore , in their latest works @xcite , taruya & sakagami interpret polytropic distributions as quasi - equilibrium states of a `` collisional '' dynamics . however , this is essentially an out - of - equilibrium result so that the connection with tsallis thermodynamics is not clear . in particular , the evolution of stellar clusters is usually described by the orbit - averaged - fokker - planck equation which exploits the timescale separation between the dynamical time and the relaxation time . it is possible that the time dependent solutions of the orbit - averaged - fokker - planck equation can be fitted by @xmath264-distributions although this equation is based on standard statistical mechanics and satisfies a h - theorem for the ordinary boltzmann entropy , not for the tsallis entropy . therefore , the results of taruya & sakagami @xcite tend to disfavor tsallis thermodynamics instead of establishing it : indeed , @xmath264-polytropes can emerge from a kinetic equation based on standard thermodynamics ( boltzmann h - theorem ) . in addition , it seems unlikely that @xmath264-polytropes are _ exact _ solutions of the orbit - averaged - fokker - planck equation , so that they represent more a convenient fit for the solutions of a complicated nonlinear time dependent partial differential equation than a fundamental outcome of a `` generalized thermodynamics '' . furthermore , it is not clear whether they attract all the solutions of the orbit - averaged - fokker - planck equation or if their `` basin of attraction '' is limited to some particular initial conditions . finally , it is not clear whether the interpretation of polytropic distributions functions given by taruya & sakagami @xcite as quasi - equilibrium states of a collisional dynamics extends to other systems with long - range interactions such as the hmf model for example . these are important questions to consider . in any case , the numerical experiments of taruya & sakagami @xcite fill the gap between the description of the `` violent relaxation '' ( first stage ) and the description of the `` gravothermal catastrophe '' ( ultimate stage ) of stellar systems . we have discussed three distinct interpretations of tsallis functional @xmath15 $ ] in astrophysics . most discussion has been devoted to vlasov - poisson systems . systems with long - range interactions spontaneously form `` coherent structures '' that are not described in general by the boltzmann distribution . this is the case for galaxies in astrophysics , vortices in 2d turbulence and quasi - stationary states in the hmf model . two types of interpretation have been given to describe these structures . in the `` generalized thermodynamical interpretation '' @xcite , the metaequilibrium states are viewed as generalized statistical equilibrium states . the maximization of the tsallis entropy @xmath15 $ ] at fixed mass and energy provides a condition of generalized thermodynamical stability . in the `` dynamical interpretation '' given here and in @xcite , the metaequilibrium states are particular stationary solutions of the vlasov equation . any df @xmath24 with @xmath25 extremizes a certain functional @xmath11=-\int c(f)d{\bf r}d{\bf v}$ ] , called a h - function , at fixed mass and energy . the condition of maximum is a criterion of nonlinear dynamical stability via the vlasov equation . these two interpretations are physically different . however , formally , the criterion of nonlinear dynamical stability for collisionless stellar systems described by the vlasov equation resembles a criterion of thermodynamical stability in the microcanonical ensemble and the criterion of nonlinear dynamical stability for barotropic stars described by the euler equation resembles a criterion of thermodynamical stability in the canonical ensemble . we have used this `` thermodynamical analogy '' to provide an original interpretation of the antonov first law in terms of ensemble inequivalence . however , the resemblance with a thermodynamical formalism is effective . in this paper , we have not discussed _ how _ a collisionless stellar system described by the vlasov - poisson system can reach a quasi - stationary distribution ( meta - equilibrium ) and what determines this metaequilibrium state . this is discussed in @xcite based on the concept of phase mixing and violent relaxation introduced by lynden - bell @xcite in 1967 . as a result of a mixing process , the coarse - grained distribution function @xmath265 achieves a stationary state on a very short timescale ( of the order of a few dynamical times ) while the fine - grained distribution function @xmath266 develops intermingled filaments at smaller and smaller scales . lynden - bell tried to predict the optimal metaequilibrium state by using statistical mechanics arguments . this prediction is complicated because we have to deal with an infinite number of constraints ( casimirs ) . therefore , the coarse - grained distribution function arising from the statistical theory of violent relaxation is non - boltzmannian in general @xcite . it can be viewed as a sort of superstatistics @xcite . in the dilute limit of his theory , lynden - bell @xcite predicted an isothermal distribution @xmath267 . he also understood that his prediction fails for high energies ( corresponding to extended orbits ) due to the complicated problem of _ incomplete relaxation _ : the systems tends to reach the statistical equilibrium state ( most mixed ) but can not attain it because the fluctuations of the gravitational field , which are the engine of the relaxation , rapidly die away . this is why other stationary solutions of the vlasov equation can emerge in practice corresponding to partially mixed states leading to confined structures . this kinetic confinement can be understood by developing dynamical models of violent relaxation @xcite . the metaequilibrium state reached by the system as a result of incomplete violent relaxation is very difficult to predict because it depends on the efficiency of mixing . it is different in general from tsallis distribution ( pure polytrope ) @xcite . for example , stellar polytropes do not provide good models of galaxies @xcite . a better model is a composite model @xcite with an isothermal core ( @xmath268 , @xmath20 ) justified by lynden - bell s theory of violent relaxation and a polytropic halo ( @xmath269 , @xmath270 ) resulting from incomplete relaxation . in this context , we have proposed to interpret the @xmath157 parameter , which can vary in space , as a _ measure of mixing _ @xcite . these general concepts also apply to 2d vortices in hydrodynamics @xcite and to the hmf model @xcite . we note that the maximization problem ( [ vh3 ] ) determining the nonlinear dynamical stability of collisionless stellar systems is consistent with the phenomenology of violent relaxation @xcite . indeed , the h - functions @xmath271=-\int c(\overline{f } ) d^{3}{\bf r}d^{3}{\bf v}$ ] calculated with the coarse - grained distribution function @xmath265 increase with time ( @xmath272 decreases ) , a property similar to the h - theorem of thermodynamics , while the energy and the mass remain approximately conserved . however , contrary to the boltzmann equation , the vlasov equation does not single out a unique h - function and the increase of the h - functions is not necessarily monotonic . because of this generalized _ selective decay principle _ @xcite , we may expect the system to achieve a metaequilibrium state which maximizes one of the h - functions ( non universal ) at fixed mass and energy . this is consistent with the criterion of nonlinear dynamical stability ( [ vh3 ] ) provided that we interpret the distribution function as the _ coarse - grained _ distribution function . it is remarkable that these two maximization principles are consistent although they have a very different physical content ( there is no notion of `` dissipation '' in the first while this lies at the heart of the second due to coarse - graining ) . plastino & plastino @xcite correctly made the link with h - functions but they also gave the impression that tsallis functional is special because of its thermodynamical properties . in addition , plastino & plastino @xcite did not explain why we have to _ maximize _ the tsallis functional . this is important because , as discussed above , this maximization problem is related to nonlinear dynamical stability , not thermodynamical stability . in particular , the notion of normalized @xmath157-averages ( as done in paper iii of @xcite ) is not relevant to settle the nonlinear dynamical stability of collisionless stellar systems . in conclusion , our discussion shows that the notions of dynamics and thermodynamics are intermingled . there are lot of analogies ( formal or real ) and differences . this is a little bit similar to the wave / particle duality in quantum mechanics , each aspect having been defended by different schools . the coherent structures that form in long - range systems have both dynamical and thermodynamical ( mixing ) properties . we think , however , that the correct description of quasi - equilibrium states is , fundamentally , in terms of nonlinear dynamical stability arguments . indeed , a collisionless stellar system can be trapped in a steady state of the vlasov equation that does _ not _ satisfy the `` thermodynamic - looking '' criterion ( [ vh3 ] ) . the maximization principle ( [ vh3 ] ) leads to df of the form @xmath24 with @xmath25 depending only on energy so that it only characterizes _ spherical _ stellar systems which seldomly occur in nature . now , there exists stationary solutions of the vlasov equation , consistent with the jeans theorem @xcite , that do not depend on the energy alone . this is the case for most observed stellar systems . in that case all the analogies with thermodynamics break down . this `` problem '' does not occur for 1d systems such as the hmf model because @xmath24 is the general form of stationary solutions of the vlasov equation for these systems . the case of stellar systems is therefore interesting with respect to 1d systems to show the limitations of the `` effective '' thermodynamical formalism . in conclusion , the dynamics and thermodynamics of systems with long - range interactions is rich and complex . we have given a lengthy discussion of these concepts in @xcite . in summary , we come to the following scenario : \1 . the rapid emergence of coherent structures in stellar dynamics ( galaxies ) , two - dimensional turbulence ( jets and vortices ) and in the hmf model ( quasi - stationary states ) can be explained by a theory of _ incomplete violent relaxation _ based on the vlasov or euler equation , see @xcite . formally , the vlasov equation is obtained when the @xmath21 limit is taken before the @xmath0 limit . \3 . due to violent relaxation and phase mixing @xcite , the coarse - grained distribution function @xmath265 converges toward a steady state @xmath275 which is a stationary solution of the vlasov equation . the convergence takes a few dynamical times , independent on @xmath1 . the fine - grained distribution @xmath266 develops filaments at smaller and smaller scales and does not achieve a steady state ( presumably ) . \4 . if mixing is complete , @xmath275 is given by lynden - bell s theory of violent relaxation @xcite . it is in general a non - standard ( non boltzmannian ) distribution depending on the initial conditions through the casimirs . it can be viewed as a form of superstatistics @xcite . if mixing is incomplete , the lynden - bell prediction fails ( by definition ! ) since @xmath275 is only partially mixed . incomplete relaxation @xcite can be explained by dynamical models of violent relaxation which involve a diffusion coefficient that vanishes in certain regions of phase - space and for large times @xcite . \6 . since @xmath275 results from a complicated mixing process , it is a nonlinearly dynamically stable stationary ( i.e. , robust ) solution of the vlasov equation . the general form of stationary solutions is given by the jeans theorem @xcite . we emphasize that , in theories of violent relaxation , we must view the steady distribution function as the _ coarse - grained _ df . it is in general very difficult to predict the state actually reached by the system in case of incomplete relaxation as it depends on the efficiency of mixing which itself depends on the `` route to equilibrium '' ( dynamics ) . it does not appear to be explained by the generalized thermodynamics introduced by tsallis in an attempt to describe non - ergodic behaviours @xcite . if @xmath276 with @xmath277 , which is a particular case of the jeans theorem , it extremizes a h - function at fixed mass and energy @xcite . furthermore , the condition of _ maximum _ is a condition of nonlinear dynamical stability @xcite . in that case , we can develop a _ thermodynamical analogy _ @xcite to investigate the nonlinear dynamical stability problem , but all the notions of thermodynamics are formal or effective . furthermore , this thermodynamical analogy can describe only spherical stellar systems with @xmath24 and @xmath25 . tsallis functional @xmath278 $ ] is the h - function corresponding to stellar polytropes which form a _ particular _ class of spherical collisionless stellar systems that are steady solutions of the vlasov equation @xcite . \9 . for @xmath0 , the system tends to relax toward a statistical equilibrium state described by the boltzmann distribution . formally , the collisional statistical equilibrium state is obtained when the @xmath0 limit is taken before the @xmath22 limit . in the case of stellar systems , the convergence to equilibrium is hampered by the escape of stars and the gravothermal catastrophe ( see the introduction ) . \10 . for intermediate collision times , the evolution of the system is described by the orbit averaged fokker - planck equation @xcite . the df passes by a sequence of quasi - equilibrium states @xmath279 which are quasi - stationary solutions of the vlasov equation slowly evolving in time due to close encounters ( finite @xmath1 effects ) . in some cases of box confined systems , this sequence is well - approximated by stellar polytropes with a time dependent index @xcite . we have also discussed another , completely independent , interpretation of tsallis functional in relation with generalized kinetic equations . in this context , tsallis functional @xmath15 $ ] is a generalized entropy taking into account `` hidden constraints '' acting on the underlying dynamics of the system . we do not reject the possibility that this generalized thermodynamical approach can be relevant to certain self - gravitating systems ( not stellar systems ) but we stress that the interpretation of tsallis functional ( sec . [ sec_genkin ] ) is then completely different from the one given for vlasov systems ( sec . [ sec_stellar ] ) @xcite . furthermore , as shown in @xcite , this generalized thermodynamical formalism ( and related kinetic theories ) can be extended to a larger class of functionals than the one proposed by tsallis .	0.5 cm we discuss different interpretations of tsallis functional in astrophysics . in principle , for @xmath0 , a self - gravitating system should reach a statistical equilibrium state described by the boltzmann distribution . however , this tendency is hampered by the escape of stars and the gravothermal catastrophe . furthermore , the relaxation time increases almost linearly with the number of particles @xmath1 so that most stellar systems are in a collisionless regime described by the vlasov equation . this equation admits an infinite number of stationary solutions . the system can be trapped in one of them as a result of phase mixing and incomplete violent relaxation and remains frozen in this quasi - stationary state for a very long time until collisional effects finally come into play . tsallis distribution functions form a particular class of stationary solutions of the vlasov equation named stellar polytropes . we interpret tsallis functional as a particular h - function in the sense of tremaine , hnon and lynden - bell [ mon . not . r. astr . soc . * 219 * , 285 ( 1986 ) ] . furthermore , we show that the criterion of nonlinear dynamical stability for spherical stellar systems described by the vlasov - poisson system resembles a criterion of thermodynamical stability in the microcanonical ensemble and that the criterion of nonlinear dynamical stability for barotropic stars described by the euler - poisson system resembles a criterion of thermodynamical stability in the canonical ensemble . accordingly , a _ thermodynamical analogy _ can be developed to investigate the nonlinear dynamical stability of barotropic stars and spherical galaxies but the notions of entropy , free energy and temperature are essentially effective . this analogy provides an interpretation of the nonlinear antonov first law in terms of ensemble inequivalence . similar ideas apply to other systems with long - range interactions such as two - dimensional vortices and the hmf model . we propose a general scenario to understand the emergence of coherent structures in long - range systems and discuss the dynamical / thermodynamical `` duality '' of their description . we stress that the thermodynamical analogy that we develop is only valid for systems whose distribution function depends only on energy . we discuss two other , independent , interpretations of tsallis functional in astrophysics , in relation with generalized kinetic equations and quasi - equilibrium states of collisional stellar systems . 0.5 cm keywords : barotropic stars ; stellar systems ; vlasov equation ; generalized thermodynamics ; nonlinear dynamical stability 0.5 cm corresponding author : p.h . chavanis ; e - mail : [email protected] ; tel : + 33 - 5 - 61558231 ; fax : + 33 - 5 - 61556065 = 10000 # 1eq.@xmath2#1 # 10=-.025em0 - 0 -0.05em0 - 0-.025em.0233em0
the continuous development of photomultiplier tubes led , in the last few years , to a major progress for the photocathodes towards cathodes with higher quantum efficiencies . for instance today , such super or ultra bialkali photomultipliers can have efficiencies of over @xmath1 ( compared to the `` old standard '' of @xmath2 ) and are now commercially available on the market . this change is particularly interesting for fluorescence telescopes , whose detection of atmospheric fluorescence light from extensive air showers has become a powerful tool for accurate measurements of the energy and the mass of ultra high energy cosmic ray particles . telescopes with mirror areas of @xmath3m@xmath0 and more allow the observation of showers out to distances of @xmath4 km . for this task it is essential to have low - noise photomultipliers which are able to detect the faint fluorescence light from the shower . the signal - to - noise ration for such a system is found to be @xmath5 where @xmath6 is the mirror area and @xmath7 the quantum efficiency of the photomultipliers in the camera . hence , higher quantum efficiencies could potentially improve the quality of the measurements and/or led to a compacter construction of telescopes . in this work we discuss important properties of such super - bialkali and ultra - bialkali photomultipliers and describe simulations of atmospheric fluorescence telescopes employing such photomultipliers . the major objectives for this simulations are the possible increase in the trigger efficiency , and thus the aperture , of the telescope and the possible better quality of the reconstruction . the task of the optical detector in fluorescence telescopes is to pick out the faint and fast signal of air showers from the ambient night - sky background noise which is of the order of 100 photons per @xmath8 in the relevant wavelength range @xcite . thus , besides the cathode and anode luminous sensitivities , also the anode dark current and after pulses are of particular importance together with the linearity and a high dynamic range . very detailed information based on large quantities of xp3062b photomultipliers @xcite have been obtained from laboratory tests reported in @xcite . such information is not yet available for super and ultra bialkali photomultipliers becoming available only very recently . thus , a smaller number of super - bialkali photomultipliers has been procured and is now under study in the laboratory , while the ultra - bialkali photomultipliers presently appear to be too expensive for being considered a realistic option for large quantities as needed in a fluorescence detector camera . first tests aim for measuring the quantum efficiency as a function of wavelength in comparison to the xp3062b photomultipliers , the homogeneity of the photocathode response , the dark current , occurrence of after - pulses , and the dynamic range . when used in combination with mug-6 filters in the aperture system , the absence of sensitivity to photons at wavelengths larger than @xmath9 becomes another issue to be verified , as the filter becomes transparent again in this wavelength range . preliminary studies show that the dark current of high quantum efficiency photomultipliers is increased compared to standard bialkali photomultipliers . this will be quantified and be reported together with other test - results at the conference . the simulations involve three different steps , their details will be explained in this section : the shower simulation , the telescope simulation and the shower reconstruction . as sample for photomultipliers with different quantum efficiencies the data of : * a photonis xp3062b , which is the photomultiplier used for the pierre auger observatory and the former hires experiment , with @xmath10 @xcite , * a hamamatsu r9420 - 100 sba ( super - bialkali ) with @xmath11 @xcite and * a hamamatsu uba ( ultra - bialkali ) photocathode with @xmath12 @xcite are used . for the second and third case only the data of the photocathodes , and not the , potentially , different gain of the multipliers , is important . this ensures an observation of effects only due to the quantum efficiency . for a fluorescence telescope the interesting part of an airshower is its longitudinal profile and the emitted fluorescence light , but not the whole particle content of the shower . for this purpose the simulation program conex @xcite was chosen . it uses a combination of monte carlo simulations and solutions of the cascade equations for airshowers and is , thus , a very fast alternative to pure monte carlo simulation programs . the primary particles were chosen to be protons with an energy distribution uniformly in @xmath13 in steps of 0.5 from @xmath14ev to @xmath15ev . for each energy step 25000 showers have been simulated with a uniform azimuth angle distribution and a zenith angle distribution according to @xmath16 between @xmath17 and @xmath18 to take into account the flat surface on which the showers will be distributed . the actual simulations of the different photomultipliers is done in this step . for convenience , we use the example of the pierre auger observatory , which has four fluorescence detector sites . to exclude effects coming from showers seen in more than one site , only one of these is simulated . for the simulations we used the so called offline framework @xcite , which is a simulation and reconstruction tool developed and used for the auger observatory . the shower cores are distributed on a @xmath19 slice in front of the telescope . the maximum distance between shower core and telescope was chosen in a way that the trigger probability ( for the standard photomultiplier ) for a shower outside this radius is lower than @xmath20 . the last step of the simulation chain is the reconstruction of the shower . this is again done by using the offline framework , which then yields all the important quantities , such as the reconstructed profile , energy and the position of the shower maximum ( @xmath21 ) . the first expectation on photomultipliers with higher quantum efficiency is the possibility to observe fainter signals , which for airshowers correlates to the possibility of observing more distant showers . one observable to measure this effect is the trigger efficiency against the distance . as trigger efficiency we define the ratio @xmath22 , the number of showers which trigger during the telescope simulation over the total number of simulated showers . this ratio is calculated for each energy - distance bin . the result is shown in figure [ triggerdistance ] . as expected , for all energies the trigger efficiency rises if the quantum efficiency of the photomultipliers becomes higher , or in other words , the telescope can see showers of @xmath23 farther away . for a clean atmosphere the expectation would be a factor of @xmath24 due to the @xmath25 intensity loss , valid for all energies . but mie and rayleigh scattering attenuate this effect for larger distances . to see the result in a more quantitative way it is useful to look at the effective areas defined as @xmath26 where the sum runs over the different distance bins with a bin - size of @xmath27 km . figure [ effarea ] shows the relative effective areas compared to the one obtained for the standard photomultipliers . the increase for all energies is obvious and becomes larger for smaller energies because of a geometrical reason , i.e. the relative area growth of a slice is of the order @xmath28 . before having a look at the reconstructed showers , the shower sample has to be cleaned up by applying several quality cuts . the main cuts are : * the reconstructed shower maximum shall lie inside the observed energy deposit profile , * the absolute uncertainty ( @xmath29 ) of the reconstructed shower maximum is smaller than @xmath30 , * the relative uncertainty of the reconstructed energy ( @xmath31 ) is smaller than @xmath32 , * the reduced @xmath33 of the gaisser - hillas fit is smaller than 5 and * the estimated amount of cherenkov light at the aperture should be smaller than @xmath34 . since the effective area increases for higher quantum efficiencies , we expect to have more `` good '' reconstructed showers in theses cases . the result in figure [ norec ] shows a significant increase of the relative number of successfully reconstructed showers . for lower energies this effect becomes larger , which is correlated to the behaviour of the effective area described in the last section . distributions of the gaisser - hillas fits . shown are the means of the distributions and their corresponding errors.,width=307 ] a good indicator for the quality of the reconstruction is the goodness of the gaisser - hillas fit , which gives the shower profile . the goodness is described by the @xmath35 of the fit . figure [ gh ] shows the means of the different @xmath35 distributions as function of the energy . as can be seen the , higher quantum efficiencies yield a significant improvement of the goodness of the fits and thus of the quality of the reconstruction . this effect is as expected because higher quantum efficiencies should yield more signals in the camera and thus facilitate the reconstruction . since the gaisser - hillas fit becomes significantly better for photomultipliers with higher quantum efficiencies , we can assume that also the resolution of the quantities derived from the profile will become better . namely , these are the energy and the position of the shower maximum ( @xmath21 ) . their mean residuals are shown in figure [ energyerr ] and figure [ xmaxerr ] , respectively . as expected they tend to lower values and thus to more precise energy and @xmath21 reconstructions . the results of the simulations done in this work clearly show that photomultipliers with higher quantum efficiencies can significantly improve the quality of airshower measurements . furthermore , they allow the observation of more distant showers or of showers with lower energies at given distance . both of this increases the aperture of the telescopes . the simulation of these new photomultipliers with high quantum efficiencies is , of course , only the first step towards a real usage in an experiment . as a next step samples of these photomultipliers are currently tested to determine their properties under conditions typically for a fluorescence telescope . a final step might then be the equipment of a whole telescope with these new photomultipliers . we gratefully acknowledge financial support of the german federal ministry of education and research ( bmbf grant 05 a0px1 ) . photonis , http://www.photonis.com/upload/industryscience/pdf/ pmt_docs / product_hexa - square.pdf k.h . et al . _ , _ qualification tests of the 11000 photomultipliers for the pierre auger observatory fluorescence detectors _ , nucl . a576 ( 2007 ) 301 hamamatsu photonics k.k . , http://jp.hamamatsu.com/resources/ products / etd / pdf / uba_sba_tpmh1305e03.pdf t. bergmann _ et al . _ , _ one - dimensional hybrid approach to extensive air shower simulation _ , astropart . ( 2007 ) 420 t. pierog _ et al . _ , _ first results of fast one - dimensional hybrid simulation of eas using conex _ , nucl . 151 ( 2006 ) 159 s. argiro _ et al . _ , _ the offline framework of the pierre auger observatory _ , nucl . a580 ( 2007 ) 1485	the detection of atmospheric fluorescence light from extensive air showers has become a powerful tool for accurate measurements of the energy and mass of ultra - high energy cosmic ray particles . employing large area imaging telescopes with mirror areas of 10m@xmath0 or more , showers out to distances of 30 km and more can be observed . matrices of low - noise photomultipliers are used to detect the faint light of the air showers against the ambient night - sky background noise . the signal - to - noise ratio of such a system is found to be proportional to the square root of the mirror area times the quantum efficiency of the phototube . thus , higher quantum efficiencies could potentially improve the quality of the measurement and/or lead to the construction of more compact telescopes . in this paper , we shall discuss such improvements to be expected from high quantum efficiency phototubes that became available on the market only very recently . a series of simulations has been performed with data of different types of commercially available high quantum efficiency phototubes . the results suggest a higher aperture and thus increased statistics for such telescopes . additionally , the quality of the reconstruction can be improved . pmts , quantum efficiency , fluorescence telescopes
the molecular - like picture as well as the single - particle picture are fundamental viewpoints to understand the structure of light nuclei.@xcite it is well known that the structure of many states in light nuclei is described successfully by the microscopic cluster model , where a group of nucleons is assumed to form a localized substructure ( cluster ) , interacting with other clusters and/or nucleons in the nucleus . the predominant cluster is the @xmath0 nucleus , which plays an important role in the cluster model , because it is the lightest and also smallest shell - closed nucleus with a binding energy as large as 28 mev , reflecting the strong four - nucleon correlation . molecular - like states in nuclei are expected to appear around the threshold energy of breakup into constituent clusters @xcite , because the intercluster binding is weak in the cluster states . for example , the ground state of @xmath8be and the second @xmath9 state of @xmath3c are known @xcite to have loosely bound @xmath10 and @xmath11 structures , respectively , which appear around the @xmath10 and @xmath11 thresholds . special attention has been paid to such four - nucleon correlations corresponding to an @xmath0-type condensation in symmetric nuclear matter , similar to the bose - einstein condensation for finite number of dilute bosonic atoms such as @xmath12rb or @xmath13na at very low temperature @xcite . several authors have discussed the possibility of @xmath0-particle condensation in low - density nuclear matter @xcite . they found that such @xmath0 condensation can occur in the low - density region below a fifth of the saturation density , although the ordinary pairing correlation can prevail at higher density . the result indicates that @xmath0 condensate states in finite nuclear system may exist in excited states of dilute density composed of weakly interacting gas of @xmath0 particles . thus , it is an interesting subject to study the structure of light nuclei from the viewpoint of @xmath0-particle condensation . recently , a new @xmath0-cluster wave function was proposed which is of the @xmath2-particle condensate type @xcite : @xmath14 where @xmath15 is the @xmath0-particle creation operator , @xmath16 denotes the internal wave function of the @xmath0 cluster , @xmath17 is the center - of - mass coordinate of the _ i_-th @xmath0 cluster , and @xmath18 presents the antisymmetrization operator among the constituent nucleons . the important characteristic in eq . ( [ wf ] ) is that the center - of - mass motion of each @xmath0 cluster is of @xmath19-wave type with the independent size parameter @xmath20 . applications of the condensate - type wave function to @xmath3c and @xmath21o showed that the second @xmath9 state of @xmath3c ( @xmath22=7.65 mev ) and fifth @xmath9 states of @xmath21o ( @xmath22=14.0 mev ) , located around the @xmath11- and @xmath23-particle thresholds , respectively , are specified by the @xmath2 condensate state , which is quite similar to the bose - einstein condensation of bosonic atoms in magnetic traps where all atoms occupy the lowest @xmath19-orbit.@xcite the calculated root - mean - square ( rms ) radius for those condensate states is about 4 fm , which is much much larger than that for the ground state ( about 2.7 fm ) . as for @xmath8be , the @xmath0-particle wave function in eqs . ( [ c_dager ] ) and ( [ wf ] ) , taking into account the axially symmetric deformation , was applied to investigate the rotational structure of the ground - band states ( @xmath9-@xmath24-@xmath25 ) @xcite . it was found that the rotational character of the @xmath8be ground state band is reproduced nicely by this wave function with deformation @xcite . the above - mentioned theoretical results for @xmath8be , @xmath3c and @xmath21o lead us to conjecture that such dilute @xmath0-cluster states near the @xmath2 threshold may also occur in other heavier @xmath1 self - conjugate nuclei . the coulomb potential barrier should play an important role to confine such dilute @xmath2-particle states , as inferred from the analyses of the @xmath8be , @xmath3c and @xmath21o nuclei . in fact , the ground state of @xmath8be , which appears at @xmath26 kev referring to the @xmath10 threshold , exists as a resonant state with very narrow width , due to the coulomb potential barrier , whose height is estimated to be @xmath27 mev . this self - trapping of @xmath0-particles by the coulomb barriers is in contrast to the case of the dilute neutral atomic condensate states , where the atoms are trapped by the external magnetic field @xcite . it is an intriguing problem and of interest to study how many @xmath0 particles can be bound in the dilute nuclear system . increasing the number of @xmath0 clusters , the rms radius of the system should become gradually larger , because the dilute character must be retained . then , the total kinetic energy of the @xmath2 system becomes significantly smaller in comparison with the potential energy , similar to the case of bosonic atoms in the condensed state . in addition , the height of the coulomb potential barrier will become steadily lower . this is due to the following reasons . the @xmath0-@xmath0 nuclear potential is short - ranged , while the coulomb potential is long - ranged . increasing the number of the @xmath0 particles , the repulsive contribution from the coulomb potentials prevails gradually over the attractive one from the @xmath0-@xmath0 nuclear potentials , because the gas - like @xmath2 system is expanding in such a way that the rms radius between two @xmath0 particles is gradually getting larger . this means that the height of the potential barrier confining the gas - like @xmath2 particles becomes lower with increasing @xmath7 , and then , there should exist a critical number ( @xmath28 ) beyond which the @xmath2 system can not sustain itself as a bound nuclear state . thus , it is interesting to estimate the number of @xmath29 as well as to study the structure of the dilute @xmath2 system up to @xmath28 . since the application of the condensate - type wave function in eqs . ( [ c_dager ] ) and ( [ wf ] ) to the general @xmath2 system is not easy and the computational limitation is @xmath30 at most , we need to study them with phenomenological models . in this paper , the gas - like @xmath2 cluster states are studied with the following two approaches : the gross - pitaevskii - equation approach and hill - wheeler - equation approach . the gross - pitaevskii equation , which is of the nonlinear - schrdinger - type , was proposed about fifty years ago to describe the single - boson motion in the dilute atomic condensate state @xcite . recent experiments starting from the middle of 1990 s have succeeded in realizing such a condensate state consisting of @xmath31-@xmath32 neutral atoms trapped by the magnetic field at very low temperature . many characteristic aspects of the dilute states are described successfully by the gross - pitaevskii equation @xcite . thus , its application to the dilute @xmath2 nuclear systems is also very promising , and we expect from such a study useful information on the condensed states . on the other hand , we here propose also a different approach . although the gross - pitaevskii equation is simple and interesting for the study of the structure of the dilute @xmath2 system , the center - of - mass motion in the @xmath2 system is not completely removed in this framework . the effect of the center - of - mass motion should be non - negligible to the total energy and rms radius etc . for small numbers of @xmath0 bosons . we , thus , formulate in this paper the framework describing the dilute @xmath2 systems free from the center - of - mass motion , with the approach of the hill - wheeler equation . the spherical @xmath2 systems as well as the deformed ones , with the axial symmetry , are discussed with this equation . the two different approaches , the gross - pitaevskii - equation approach and the hill - wheeler - equation approach , are complementary to one another , and we will obtain useful understanding of the dilute multi @xmath0-particle states . starting from @xmath8be , the structure of the gas - like @xmath2 boson systems with @xmath5 is investigated and the critical number @xmath29 is estimated with those two different approaches . the paper is organized as follows . we formulate in sec . ii the gross - pitaevskii - equation approach and the hill - wheeler - equation approach for the dilute @xmath2 system . in sec . iii the calculated results are presented , and the characteristics of the dilute @xmath2 states are discussed with emphasize on the @xmath7 dependence of their energies and rms radii . a summary finally is given in sec . in this section , we formulate the two approaches to study the structure of the dilute @xmath2 nuclear systems : the gross - pitaevskii - equation approach and hill - wheeler - equation approach . in the mean field approach , the total wave function of the condensate @xmath2-boson system is represented as @xmath33 where @xmath34 and @xmath17 are the normalized single-@xmath0 wave function and coordinate of the _ i_-th @xmath0 boson . then , the equation of motion for the @xmath0 boson , called as the gross - pitavskii equation , is of non - linear schrdinger - type , @xmath35 where @xmath36 stands for the mass of the @xmath0 particle , @xmath37 is the mean - field potential of @xmath0-particles , and @xmath38 ( @xmath39 ) denotes the @xmath10 ( @xmath11 ) interaction . the center - of - mass kinetic energy correction , @xmath40 , is taken into account together with the finite number corrections , @xmath41 , etc . in the present study , only the @xmath19-wave state is solved self - consistently with the iterative method . the total energy of the @xmath2 system @xmath42 is presented as @xmath43,\\ & & { \langle t\rangle}=\left(1-\frac{1}{n}\right)\times{\langle \varphi(\vec{r } ) \|-\frac{\hbar^2}{2m}\vec{\nabla}^2\| \varphi(\vec{r})\rangle},\\ & & { \langle \upsilon_2\rangle}={\langle \varphi(\vec{r})\varphi(\vec{r}')\|\upsilon_2(\vec{r},\vec{r}')\| \varphi(\vec{r})\varphi(\vec{r}')\rangle},\\ & & { \langle \upsilon_3\rangle}={\langle \varphi(\vec{r})\varphi(\vec{r}')\varphi(\vec{r}'')\| \upsilon_3(\vec{r},\vec{r}',\vec{r}'')\| \varphi(\vec{r})\varphi(\vec{r}')\varphi(\vec{r}'')\rangle},\end{aligned}\ ] ] and the eigen energy @xmath44 in eq . ( [ gpe ] ) is given as @xmath45 the nuclear rms radius in the @xmath2 state is evaluated as , @xmath46 where we take into account the finite size effect of the @xmath0 particle and the correction of the center - of - mass motion . the framework of the gross - pitaevskii approach is simple and useful for the study of the structure of the dilute @xmath2 systems . the center - of - mass motion , however , is not removed exactly in this approach . the effect on the total energy and rms radius etc . is not negligible for the small - number @xmath0-boson systems . we formulate here the hill - wheeler - equation approach for the axially symmetric @xmath2 systems as well as for the spherical ones , in which the center - of - mass motion is completely eliminated . only the hill - wheeler equation for the deformed @xmath2 systems with the axial symmetry is provided in this section , because the limit of the spherical systems is included in the deformed case . the single @xmath0-particle wave function should be mainly in the lowest @xmath19 state in the gas - like @xmath2 boson system . thus , the model wave function of the axially symmetric @xmath2 system should be described in terms of a superposition of a gaussian basis with axially symmetric deformation ( taking the @xmath47-axis as the symmetric axis ) as follows , @xmath48 where @xmath17 ( @xmath49 ) denotes the coordinate of the _ i_-th @xmath0 boson ( the center - of - mass coordinate of the @xmath2 system ) , and @xmath50 ( @xmath51 ) presents the gaussian size parameter of the @xmath52 and @xmath53 directions ( @xmath47 ) . the wave function in eq . ( [ total_wf ] ) is totally symmetric under any exchange of two @xmath0 bosons . the angular - momentum - projected total wave function for dilute @xmath2 systems free from the center - of - mass motion is given as @xmath54 where @xmath55 denotes the rotation operator around the _ y _ axis . setting @xmath56 in eq . ( [ projected_wf ] ) , we obtain the wave function for the spherical @xmath2 system , where only @xmath5 state is allowed . the matrix element of a translational invariant scalar operator @xmath57 with respect to the angular momentum projected @xmath2 wave function in eq . ( [ j_projected_basis ] ) is evaluated as @xmath58 the total hamiltonian of the @xmath2-boson system is given as @xmath59 where @xmath38 and @xmath39 denotes , respectively , the @xmath10 and @xmath11 interactions . the kinetic energy of the center - of - mass motion ( @xmath60 ) is subtracted from the hamiltonian . the equation of motion for the dilute @xmath2-boson states are given in terms of the hill - wheeler equation @xcite : @xmath61 where @xmath62 is the total hamiltonian of the @xmath2-boson system in eq . ( [ hw_hamiltonian ] ) . the coefficients @xmath63 and eigen energies @xmath64 are obtained by solving the hill - wheeler equation . in the numerical calculation , the gaussian size parameters @xmath50 and @xmath51 are discretized and chosen to be of geometric progression , @xmath65 the above choice of the gaussian range parameters is found to be suitable for describing the dilute @xmath2 states . the nuclear rms radius measured from the center - of - mass coordinate in the @xmath2 state , is expressed as , @xmath66 where we take into account the finite size effect of the @xmath0 particle . the rms distance between two @xmath0 particles is given as @xmath67 thus , it is proportional to the rms radius of an @xmath0-particle from the center - of - mass coordinate . in the present paper , we use two kinds of effective potentials : the density - dependent potential and phenomenological @xmath10 plus @xmath11 potential . they are applied to the gross - pitaevskii equation and the hill - wheeler equation for the study of the structure of dilute @xmath2 states . the density dependent potential consists of the gaussian - type @xmath0-@xmath0 potential including a density - dependent term , which is of similar form as the gogny potential ( known as an effective @xmath68 potential ) used in nuclear mean - field calculations , @xmath69 -130\exp\left[-0.475 ^ 2(\vec{r}-\vec{r}')^2\right]\nonumber\\ & & \hspace*{1.5cm}+(4\pi)^2g\delta(\vec{r}-\vec{r}')\rho\left(\frac{\vec{r}+\vec{r}'}{2}\right ) + \upsilon_{coul}(\vec{r},\vec{r } ' ) , \label{dd_pot } \end{aligned}\ ] ] where the units of @xmath38 and @xmath70 are mev and fm , respectively , and @xmath71 denotes the density of the @xmath2 system . the folded coulomb potential @xmath72 is presented as @xmath73 the gaussian - potential part in eq . ( [ dd_pot ] ) is based on the ali - bodmer potential @xcite , which is known to reproduce well the elastic @xmath0-@xmath0 scattering phase shift up to about 60 mev for @xmath74 mev . on the contrary , here , the two parameters , @xmath75 and @xmath76 , are chosen so as to reproduce well the experimental energy ( @xmath77=0.38 mev ) and the calculated rms radius ( 4.29 fm by tohsaki et al . @xcite ) for the @xmath78 state of @xmath3c by solving the gross - pitaevskii equation in eq . ( [ gpe ] ) . the results are @xmath75=271 mev and @xmath76=1650 mev@xmath79@xmath80 , where the calculated energy and rms radius for the @xmath11 system are 0.38 mev and 4.14 fm , respectively , which are discussed in sec . although the choice of the phenomenological potential in eq . ( [ dd_pot ] ) is rather rough , it is interesting to study systematically the structure of the @xmath2 condensate states as a function of @xmath7 . there are many phenomenological @xmath10 potentials proposed so far . since the @xmath0 particle is treated as a point - like boson in the present study , we will use a @xmath10 potential taking into account the pauli blocking effect . the typical potential is the ali - bodmer one @xcite , which is used frequently in the structure calculation ; @xmath81 -130 \exp\left[-0.475 ^ 2(\vec{r}-\vec{r}')^2\right ] + \upsilon_{coul}(\vec{r},\vec{r}'),$ ] where @xmath72 denotes the folded coulomb potential given in eq . ( [ coulomb_pot ] ) . the strong repulsion in the inner region prevents the @xmath10 particles from approaching one another . it is , however , found that the potential is not suitable to describe the property of the compact shell - model - like structure of the @xmath3c ground state with the @xmath11 boson model @xcite . however , we may use it for the dilute @xmath2 states . the ali - bodmer potential , however , has the following three unfavorable properties for the present calculation . first is that the potential does not give the experimental resonant energy of the @xmath8be ground state ( @xmath82=68 kev vs. @xmath83=92 kev ) , although the @xmath0-@xmath0 scattering phase shift is reproduced nicely up to about 60 mev . the second is that applying the potential to the @xmath11 boson system the lowest energy state obtained corresponds to a relatively compact @xmath11-structure state , although the condensate state appears around the @xmath11 threshold . according to the stochastic variational calculation @xcite , for example , the calculated energy and rms radius are , respectively , @xmath84 mev and @xmath85=3.15 fm @xcite . the results indicate that the ali - bodmer potential is not adequate for describing the dilute @xmath11-boson state of @xmath3c . the third unfavorable point comes from the fact that the strong repulsive character in the short - range region of the ali - bodmer potential ( @xmath86 mev ) leads us to treat exactly the short - range correlation between the @xmath10 bosons . the treatment is difficult and needs time - consuming numerical calculations for solving general @xmath2-boson systems , even if employing modern numerical methods for many body systems . we construct here an effective @xmath10 potential with a weak repulsive part ( soft core ) , which overcomes the above - mentioned three unfavorable properties . the use of such a soft - core - type @xmath10 potential is suitable in the present study , because we discuss the gas - like @xmath2 states , where the pauli blocking effect between two @xmath0 particles is considerably weakened . the important point for the determination of the potential parameters , other than the condition of reproducing the experimental resonant energy for the @xmath8be ground state , is that the @xmath0-@xmath0 wave function for the resonant state , should have a loosely - bound structure of the two @xmath0 particles . according to many structure calculations of @xmath8be , the amplitude of the radial part of the @xmath0-@xmath0 relative wave function must be small in the inner region and have a maximum value around @xmath87 fm ( where @xmath88 is the relative coordinate between the two @xmath0 particles ) . this condition ensures that the ground state of @xmath8be has a dilute @xmath10 structure . with a careful search of the potential parameters , we determined the effective @xmath10 potential as follows , @xmath89 -34.101\exp\left[-0.3 ^ 2(\vec{r}-\vec{r}')^2\right ] + \upsilon_{coul}(\vec{r},\vec{r } ' ) , \label{2_body}\end{aligned}\ ] ] where the units of @xmath38 and @xmath70 are mev and fm , respectively , and @xmath72 is the folded coulomb potential . the calculated resonant energy of @xmath8be is @xmath90=92 kev , in agreement with the experimental data ( @xmath83=92 kev ) . figure [ fig:1 ] shows the radial part of the relative wave function between the @xmath10 clusters for the resonant state . we see that the amplitude is relatively small for @xmath91 fm , and has a maximum value around @xmath70=4 fm . in spite of the fact that the ali - bodmer potential gives a wave function which is almost zero for @xmath92 fm , the small but finite amplitude around @xmath91 fm should hardly give any effect for dilute multi @xmath0 cluster states . applying this effective potential to the @xmath11- and @xmath23-boson systems indicates that we get the desired dilute @xmath11- and @xmath23-structure states for @xmath3c and @xmath21o , respectively , within our framework , as shown below . in the present study , we introduce the phenomenological @xmath11 potential ( @xmath93 ) with repulsive character , as given in ref . @xcite , @xmath94}\right\}},\label{3_body}\end{aligned}\ ] ] where the units of @xmath93 and @xmath70 are mev and fm , respectively . this potential has been used in the @xmath11 and @xmath23 orthogonally condition model ( ocm @xcite ) for calculations of @xmath3c and @xmath21o so as to reproduce the ground - state energies with respect to the @xmath11 and @xmath23 thresholds , respectively @xcite . in the model , the pauli principle is taken into account in the relative wave function between two @xmath0 particles , and a deep attractive potential is used for the @xmath0-@xmath0 potential . thus , the ocm is able to describe not only the shell - model - like compact states but also the dilute gas - like states . according to the results , the repulsive @xmath11 potential gives a large effect to the ground - state energies of @xmath3c and @xmath21o with the compact @xmath2 structure , while its effect is very small for dilute @xmath11 and @xmath23 states . a non - negligible effect , however , can be expected in large - number @xmath2 dilute systems , if we take into account the fact that the contribution of the binding energy from the @xmath11 potential , proportional to @xmath95 , raises strongly with increasing @xmath7 . the reason of why we introduce the repulsive @xmath11 potential is given as follows . let us define here the total kinetic energy and two - body potential energy in the dilute @xmath2 system as @xmath96 and @xmath97 , respectively . in case of the @xmath8be ground state , the experimental energy is @xmath98 mev with respect to the @xmath10 threshold , where @xmath99 ( @xmath100 ) is positive ( negative ) . for an arbitrary dilute @xmath2-boson system , the total kinetic energy and two - body potential energy are given as @xmath101 and @xmath102 , respectively , where the center - of - mass kinetic energy is subtracted . increasing the number of the @xmath0 particles , therefore , the potential energy prevails over the kinetic energy , and then , the system falls gradually into a collapsed state . this indicates that something like a density - dependent force with the repulsive character is needed to avoid the collapse in the large - number @xmath2-boson system . the present repulsive @xmath11 potential in eq . ( [ 3_body ] ) plays a role similar to the density - dependent force . on the other hand , the density - dependent potential given in eq . ( [ dd_pot ] ) is also used when solving the gross - pitaevskii equation and the hill - wheeler equation . the reason of why the density - dependent potential is introduced there is the same as that discussed here . the gross - pitaevskii equation is solved with the two different types of effective @xmath0-@xmath0 potentials : 1 ) the density - dependent potential [ see eq . ( [ dd_pot ] ) ] and 2 ) the phenomenological @xmath10 potential with the @xmath11 potential [ see eqs . ( [ 2_body ] ) and ( [ 3_body ] ) ] . first of all , we will discuss the results with the density - dependent potential and then those with the phenomenological potentials . the calculated total energies of the @xmath2 systems measured from the @xmath2 threshold are demonstrated in fig . [ fig:2 ] as well as the calculated nuclear rms radii defined in eq . ( [ rms_n ] ) , where the density - dependent @xmath0-@xmath0 potential is used . the total energy and the rms radius are getting larger with increasing @xmath7 . this means that the system is expanding steadily with increase of @xmath7 . in comparison with the rms radius of the ground state of each nucleus with the empirical formula ( @xmath103 fm ) , the results from the gross - pitaevskii equation are much larger than those for the ground states . thus , the states obtained here can be identified with the dilute @xmath2 states . it is noted that they are obtained naturally from the gross - pitaevskii equation with the density - dependent @xmath0-@xmath0 potential , whose parameters were determined so as to reproduce well the experimental energy ( @xmath104=0.38 mev ) and the calculated rms radius by tohsaki et al . ( 4.29 fm ) @xcite for the @xmath105 state of @xmath3c . in order to study the structure of the dilute @xmath2 states , it is instructive to see the single @xmath0 potential defined in eq . ( [ gpe_alpha_pot ] ) . figures [ fig:3](a)@xmath106(j ) show the ones for the @xmath107 systems . let us first discuss the @xmath11 and @xmath23 cases . the remarkable characteristics of the potentials can be presented as follows : 1 ) the almost flat behavior of the potential in the inside region , and 2 ) the coulomb potential barrier in the outer region . the appearance of the flat potential region is very impressive , if one takes into account the fact that the two - range gaussian term ( attractive ) , density - dependent term ( repulsive ) and the coulomb - potential term ( repulsive ) contribute significantly to the single @xmath0 potential in the inside region [ see figs . [ fig:3](a ) and ( b ) ] . table [ tab:1 ] shows the calculated single @xmath0 particle energy and contributions from the kinetic energy , two - range gaussian term in eq . ( [ dd_pot ] ) , density - dependent term in eq . ( [ dd_pot ] ) and the coulomb potential . it is found that the kinetic energy in the @xmath11 and @xmath23 systems is not negligible but small in comparison with the two - range - gaussian term and/or coulomb potential energy . this indicates that the thomas - fermi approximation , neglecting the kinetic energy term , is roughly realized in the system . consequently , there appears the flat potential region in the single @xmath0 potential , whose behavior is similar to the dilute atomic condensate state trapped by the magnetic fields at very low temperature @xcite . on the other hand , the appearance of the coulomb potential barrier plays an important role in confining the @xmath0 bosons in the inside region . it is noted that the barrier comes out naturally from the self - consistent calculation of the gross - pitaevskii equation . the single @xmath0 particle energy for the @xmath11 and @xmath23 systems are smaller than the coulomb potential barrier . this means that the dilute states are quasi - stable against @xmath0 decay . increasing the number of @xmath0 bosons beyond @xmath108 , the following interesting features can be seen in the single @xmath0 particle potentials : 1 ) the depth of the flat potential becomes shallower , and its range is expanding to the outer region faster than @xmath109 , and 2 ) the height of the coulomb potential barrier is getting lower and almost disappears at around @xmath110 . the first behavior of the potentials means that the @xmath2 system is inflating with increase of @xmath7 , as inferred from the behavior of the rms radius shown in fig . [ fig:2 ] . the reason of why the depth of the potential becomes shallower is related to the fact that the single @xmath0 particle energy is getting larger . it is given as follows : from table [ tab:1 ] , the contribution from the kinetic energy to the single @xmath0 particle energy becomes smaller with increase of @xmath7 , reflecting the inflation of the dilute @xmath2 system . this fact indicates that the thomas - fermi approximation is getting better , and then , the single @xmath0 particle energy is given approximately as the contribution from only the potential energies . the potential energies consist of three contributions , namely , the two - range gaussian term , density - dependent term , and coulomb potential [ see eq . ( [ dd_pot ] ) ] , where the first and second ones are short - range , and the third is long - range . when the dilute @xmath2 system is inflating with @xmath7 and the distance between two @xmath0 bosons is getting larger , the contribution from the long - range potential should overcome steadily the one from the short - range one . consequently , the potential depth ( single @xmath0 particle energy ) becomes gradually shallower ( larger ) and the height of the coulomb barrier is getting lower , with increase of @xmath7 . the increase of the single @xmath0 particle energy means that the total energy of the dilute @xmath2 system becomes larger with @xmath7 . the present results are consistent with the behavior of the @xmath7 dependence of the total energy ( see fig . [ fig:2 ] ) . the second interesting behavior of the single @xmath0 particle potential with @xmath111 is that there exists a critical number of @xmath0 bosons , @xmath29 , beyond which the system is not confined anymore , as mentioned in sec . i. around @xmath110 , the coulomb barrier has almost disappeared , and the single @xmath0 particle potential is nearly flat up to the outer region , although the most outer region , dominated by only the coulomb potential , is falling to zero ( not illustrated in fig . [ fig:3 ] ) . thus , we can roughly estimate the critical number as @xmath6 , namely , @xmath4ca . it is interesting to compare the above results with those obtained by using the gross - pitaevskii approach with the phenomenological @xmath10 and @xmath11 potentials given in eqs . ( [ 2_body ] ) and ( [ 3_body ] ) , respectively . the calculated total energies and rms radii for @xmath2 systems are shown in fig . [ fig:4 ] . we find that they are in good agreement with those of the case with the density - dependent potential in the gross - pitaevskii equation within about 10% except for the @xmath112 and @xmath113 cases . table [ tab:2 ] shows the single @xmath0 particle energies and contributions from the kinetic energy and potential energies . comparing them with those in table [ tab:1 ] , the @xmath7 dependence of the kinetic energy , coulomb potential energy and the sum of other potential energies is rather similar to the one with the density - dependent potential in the gross - pitaevskii equation except for the @xmath112 and @xmath113 systems . this is surprising if taking into account the fact that the forms of the two kinds of the effective @xmath0-@xmath0 potentials are quite different from one to another . these results might indicate that the dilute @xmath2 systems with @xmath111 do not depend strongly on the details of the effective @xmath0-@xmath0 potential , while those with a small - number-@xmath2 ( @xmath112 and 4 ) are sensitive to the potential . the single @xmath0 particle potentials are shown in fig . [ fig:5 ] . for the @xmath11 and @xmath23 systems , we see the almost flat potential behavior in the inside region , and the coulomb potential barrier in the outer region , while increasing the number of @xmath0 bosons from @xmath108 , the depth of the flat potential becomes shallower , and its range is expanding to the outer region , and the height of the coulomb potential barrier is getting lower and almost disappears at around @xmath110 . the qualitative potential behaviors are almost the same as those in case of the density - dependent potential with the gross - pitaevskii equation ( see fig . [ fig:3 ] ) . in fact , the behavior of the single @xmath0 particle potential in fig . [ fig:5 ] is in good agreement with that in fig . [ fig:3 ] within about 10 % . on the other hand , we can conjecture the critical number @xmath29 from the behavior of the single @xmath0 potentials in fig . [ fig:5 ] , which is estimated roughly as @xmath6 , the result being the same as that of the case with the density - dependent potential . the gross - pitaevskii equation is simple and useful for the study of the structure of the dilute @xmath2 system , as seen in the previous subsection . the center - of - mass motion in the @xmath2 system , however , is not completely removed in this framework . the hill - wheeler - equation approach is free from the center - of - mass motion of the @xmath2 system . the effect of the center - of - mass motion is expected to be non - negligible for the total energy and rms radius etc . in small - number @xmath0-boson systems , in particular , @xmath112 and 4 . thus , it is important to study the dilute @xmath2 systems with the use of the hill - wheeler equation . the two approaches should give the same results for dilute @xmath2 systems with rather large number of @xmath7 . in this subsection , we first demonstrate that the hill - wheeler - equation approach is useful to describe the dilute @xmath2 states with the spherical shape , and the results are briefly compared with those from the gross - pitaevskii equation . then , the approach is applied to the deformed @xmath2 systems ( @xmath5 ) with axial symmetry . we use the phenomenological @xmath10 and @xmath11 potentials of eqs . ( [ 2_body ] ) and ( [ 3_body ] ) as the effective @xmath114 potential . the calculated total energies and rms radii for spherical @xmath2 states are illustrated in fig . [ fig:6 ] , in which we show those obtained by solving the gross - pitaevskii equation with the same effective @xmath0-@xmath0 potential for comparison . we found that the calculated total energies in the hill - wheeler approach are almost the same as those in the gross - pitaevskii approach . this indicates that the center - of - mass kinetic energy correction in eq . ( [ gpe ] ) is a very good approximation to remove the effect in the total energy . the non - negligible deviation between the two frameworks , however , can be seen in the rms radii , in particular , we have about 20 % deviation for the @xmath11 system , although that is getting smaller with increasing @xmath7 and it is almost zero in the @xmath115 systems ( the discrepancy in the @xmath116 and @xmath117 systems will be discussed later ) . the results tell us that the center - of - mass correction for the rms radius in the gross - pitaevskii - equation approach [ see eqs . ( [ rms_n ] ) and ( [ rms_n_cor ] ) ] is not very good for the @xmath11 and @xmath23 systems , while it is relatively good for the @xmath118 systems , and the difference between the two approaches is less than about 10 % , and it is diminishing with increasing @xmath7 . in fig . [ fig:6 ] we see that the deviation of the rms radius appears again in the @xmath116 and @xmath117 systems . the reason of why this deviation occurs can be understood , recalling the following facts : according to the results of the gross - pitaevskii equation with the phenomenological @xmath10 and @xmath11 potentials , the @xmath116 system is the critical one which can not exist as a nuclear state , as mentioned in the previous section . in such a critical system , the behavior of the wave functions in the outer ( inner ) region is very sensitive ( not very sensitive ) to how to solve the equations and to obtain the wavefunctions . the value of the rms radius ( total energy ) is generally sensitive ( not very sensitive ) to the behavior of the wave function in the outer region . thus , the re - occurrence of the deviation in the rms radii indicates that the @xmath116 system is critical in the hill - wheeler - equation approach . the above results show us that the hill - wheeler - equation approach is very useful to describe the dilute @xmath2 states as well as the gross - pitaevskii - equation approach . thus , we can apply the approach to the deformed @xmath2 system with the axial deformation . the @xmath2 states ( @xmath5 ) with the axial deformation are obtained by solving the hill - wheeler equation in eq . ( [ hill_wheeler ] ) with the gaussian size parameter set a given in table [ tab:3 ] . figure [ fig:7 ] illustrates the calculated energies , nuclear rms radii defined in eq . ( [ rms_n ] ) and the rms distances between @xmath10 bosons . the detailed values are shown in table [ tab:4 ] . the rms distance between @xmath10 bosons ( @xmath119 fm ) for @xmath120 is considerably larger than that of the @xmath8be ground state ( @xmath121 fm ) . the result indicates that the @xmath2 states obtained here are of very dilute @xmath2 structure . the total energy of the @xmath2 state increases gradually with @xmath7 , although those for @xmath7=3 and 4 are not changed very much . the latter is in contrast to the results of the spherical case together with those for the gross - pitaevskii - equation approach ( see figs . [ fig:2 ] , [ fig:4 ] and [ fig:6 ] ) . in the @xmath11 and @xmath23 systems , the total energies and nuclear rms radii are respectively given as follows ; @xmath122 and 0.13 mev , and @xmath85=3.73 and 3.90 fm . the values are in good correspondence to those for the condensate @xmath0 cluster states discussed by tohsaki et al . @xcite , where their calculated results are @xmath123=0.5 and @xmath124 mev ( vs. @xmath125=0.38 and @xmath126 mev ) and @xmath85=4.29 and 3.97 fm , respectively , for the dilute @xmath11 and @xmath23 states . it is instructive here to compare the present results with those in the spherical @xmath2 states in order to estimate the effect of deformation . the comparison of the total energy and rms radius is given in table [ tab:4 ] . the energy gains ( reductions of rms radius ) due to the deformation in the @xmath11 , @xmath23 and @xmath127 systems are , respectively , 0.6 , 1.5 and 1.7 mev ( 11 % , 18 % and 19 % ) , while this is getting smaller for @xmath128 . thus , we found that the deformation effect is significant for relatively small - number @xmath0 systems , and shows a good correspondence of our results for the @xmath11 and @xmath23 states with those by tohsaki et al . the present result that the dilute @xmath2 states ( @xmath5 ) may be deformed is natural if taking into account the fact that a gas - like @xmath2 state with relatively small number can easily be deformed . if it is right , there may exist dilute @xmath2 nuclear states with @xmath129 and @xmath25 etc . in fact , a candidate of the dilute @xmath11 state with @xmath129 is observed at @xmath64=3.3 mev measured from the @xmath11 threshold @xcite . the dilute multi @xmath0 cluster states with non - zero angular momentum will be discussed elsewhere . from fig . [ fig:7](a ) , we notice that the gas - like @xmath2 states with @xmath130 appear above the @xmath2 threshold , not close to it , in contrast to the fact that the dilute @xmath11 and @xmath23 states are located in the vicinity of their respective thresholds . this is a non - trivial result of the present calculation . the reason of why the energy of the dilute @xmath2 states increases with @xmath7 is given as follows . let us consider the effective energy of the @xmath131 system in the dilute @xmath2 state , which corresponds to the quantity of the total energy of the @xmath2 state divided by the number of @xmath0 pairs which is given in table [ tab:4 ] . since the dilute character allows us to neglect approximately the kinetic energy , the effective energy of the @xmath131 system could be given mainly as the sum of the @xmath10 nuclear potential energy and its coulomb potential energy ( the contribution from the @xmath11 potential may be neglected approximately because of the very small amount ) . figure [ fig:8 ] shows the @xmath10 potential used in the present study , where we also show the @xmath10 nuclear potential [ see eq . ( [ 2_body ] ) ] and its coulomb potential [ see eq . ( [ coulomb_pot ] ) ] . the attraction of the former is largest around 4 fm and is negligible beyond about 7 fm , while the coulomb potential is long - ranged and its repulsion is substantial even beyond 7 fm . increasing the number of the @xmath0 particles , the gas - like @xmath2 system is expanding and the rms distance between two @xmath0 particles is also becoming larger ( @xmath132 fm ) in the case of @xmath130 , as shown in fig . [ fig:7 ] . then , the attractive contribution from the @xmath10 nuclear potential to the effective @xmath131 energy should become noticeably smaller with increasing @xmath7 , reflecting the short - range ( attractive ) behavior of the potential . the repulsive contribution from the coulomb potential , however , should not be much smaller than in the case of the @xmath10 nuclear potential , and dominate over the contribution from the nuclear potential , because of the long - range ( repulsive ) behavior . these facts explain the raise of the effective @xmath131 energy with @xmath7 , which means that the energy of the dilute @xmath2 states increases with @xmath7 and does not stay around the threshold energy . in table [ tab:4 ] we can find quantitatively the increase of the effective @xmath10 system with @xmath7 . the gradual dominance of the total coulomb potential energy over the total @xmath10 nuclear potential energy in the @xmath2 system can be also seen in table [ tab:4 ] . it is interesting to see the role of the @xmath11 potentials in the dilute @xmath2 system . from table [ tab:4 ] , the contribution to the total energy is small in the case of @xmath3c and @xmath21o , while it gives a non - negligible effect with increasing @xmath7 , as seen in table [ tab:3 ] , because it is proportional to the number of @xmath95 . in fact , the energy @xmath133 becomes steadily larger with the number of @xmath0 bosons , although the quantity of the energy divided by the number of trios of the @xmath0 bosons is decreasing with @xmath7 , reflecting the fact that the @xmath2 system is expanding with @xmath7 . it is instructive to study the case of no @xmath11 potentials within the present framework . the calculated results show that the @xmath2 system gradually falls into a collapsed state with increasing @xmath7 as discussed before ; for example , @xmath134 mev and @xmath135 fm for @xmath136 . the result tells us that the dilute @xmath0 boson states appear under the cooperative work between the two - body and three - body potentials in the present model . finally , we estimate the critical number @xmath29 of the dilute @xmath2 state beyond which the system is unbound . since the critical state should be very unstable , it is difficult to determine it exactly with the present hill - wheeler - equation approach . however , we can deduce it approximately from studying the stability of the calculated eigen energies and rms radii against changing the model space determined by the gaussian size parameters in eqs . ( [ gs_para_1 ] ) and ( [ gs_para_3 ] ) . the procedure is as follows . the present framework is able to describe dilute states trapped by the coulomb barrier under the condition that we choose the range of the gaussian size parameters wide enough to cover a configuration space over the whole rms radius of the states . if the state is stable or relatively stable , the calculated energies and rms radii are not changed very much against the variation of the gaussian size parameters . otherwise , those depend sensitively on the choice of the parameters . using the wide model space , however , a special attention must be paid to investigate the eigen - states obtained from the hill - wheeler equation , because a discretized continuum state , which is unphysical , has a chance to become the lowest state in energy in the present variational calculation . we can easily identify it by investigating the behavior of its energy and/or rms radius against changing the gaussian size parameters . the unphysical state usually has an abnormally long rms radius , which is a little smaller than or almost the same as the value of @xmath137 or @xmath138 in eqs . ( [ gs_para_1 ] ) and ( [ gs_para_3 ] ) . table [ tab:5 ] shows the calculated results for the two parameter sets a and b ( see table [ tab:3 ] ) . the calculated energies and rms radii for the dilute @xmath2 states with @xmath139 are almost the same for both the parameter sets a and b , while we see some discrepancies in those with @xmath140 , in particular , in the rms radii , for the two parameter sets . even if we use other gaussian size parameter sets which cover the configuration space over the rms radius for the dilute states , the results obtained are found to be similar to those mentioned above . thus , the critical number of the dilute @xmath2 boson systems is estimated roughly to be @xmath6 , namely , @xmath4ca , in the present study . it is noted that the number is the same as that in the spherical case as well as in the gross - pitaevskii - equation approach . the dilute @xmath2 cluster condensate states with @xmath5 have been studied with the gross - pitaevskii equation ( gp ) and the hill - wheeler equation ( hw ) , where the @xmath0 cluster is treated as a structureless boson . two kinds of effective @xmath0-@xmath0 potentials were used : the density - dependent @xmath0-@xmath0 potential ( dd ) and the phenomenological @xmath10 potential plus 3-body @xmath0 potential ( pp ) , and we also included the folded coulomb potential between the @xmath10 bosons . both potentials ( only the latter ) were ( was ) applied to study the structure of the spherical ( spherical and axially deformed ) dilute @xmath2 states with the gross - pitaevskii equation ( hill - wheeler equation ) . thus , we have studied the dilute multi @xmath0 cluster states with the four different frameworks ; ( equation , potential , shape)=(gp , dd , s ) , ( gp , pp , s ) , ( hw , pp , s ) and ( hw , pp , d ) , where s and d denotes the spherical and deformed shapes , respectively . the main results to be emphasized here are as follows : 1 ) all of the @xmath2 states obtained show the dilute @xmath2 structure . they are common to all of the four different frameworks . the reason of why the total energy of the gas - like @xmath2 state increases gradually with @xmath7 and do not remain around their @xmath2 threshold values is understood as the competition between the nuclear @xmath0-@xmath0 nuclear potential ( attractive ) and its coulomb potential ( repulsive ) . in fact , increasing the number of @xmath7 , the @xmath0-@xmath0 distance is becoming larger ( @xmath141 fm for @xmath120 ) , and then the contribution from the @xmath0-@xmath0 nuclear potential per @xmath10 pair in the dilute @xmath2 state is decreasing rapidly because of the short - range character , while that from the coulomb one is decreasing very slowly and remains almost constant for @xmath142 , reflecting the @xmath143 character of the coulomb potential . 2 ) the @xmath7 dependence of the behavior of the calculated single @xmath0 potentials obtained by solving the gross - pitaevskii equation within the ( gs , dd , s ) and ( gw , pp , s ) frameworks is impressive . for the @xmath11 and @xmath23 systems , we see an almost flat potential behavior in the inside region , and the coulomb potential barrier in the outer region , while increasing the number of @xmath0 bosons from @xmath108 , the depth of the flat potential becomes shallower , and its range is expanding to the outer region , while the height of the coulomb potential barrier is decreasing and almost disappears at around @xmath110 . the origin of the appearance of the flat region is mainly due to the validity of the thomas - fermi approximation , namely , the neglect of the contribution from kinetic energy because of the dilute character of the @xmath2 system . the result is analogous to the dilute atomic condensate state trapped by the magnetic fields at very low temperature . on the other hand , the results for the total energy and rms radius were quite similar for the above - mentioned two frameworks except for @xmath7=3 and 4 . the result might indicate that the dilute @xmath2 states with @xmath111 are not very sensitive to the details of the effective @xmath0-@xmath0 potential ( dd or pp ) . 3 ) comparing the results from the gross - pitaevskii equation ( gp ) and hill - wheeler equation ( hw ) , we could see the effect of the neglect of the center - of - mass motion in the gross - pitaevskii equation and also the usefulness of the hill - wheeler - equation approach for describing the dilute @xmath2 system . in the small - number @xmath2 systems , the effect was found to be non - negligible , in particular , for the rms radius , while there is less than about 10 % deviation for the @xmath118 systems . 4 ) the axial deformation effect in the dilute @xmath2 systems is substantial for the small - number @xmath0 systems , but it is getting smaller for @xmath128 . in fact , the energy gain and reduction of the rms radius due to the deformation are @xmath144 mev and @xmath145 % , respectively , for the former system . due to the effect , the calculated results of the total energy and rms radius for @xmath7=3 and 4 in the ( hw , pp , d ) framework are improved from those in the ( hw , pp , s ) framework , and give a good correspondence with those for the condensate @xmath0 cluster states discussed by tohsaki et al . @xcite the present result that the dilute @xmath2 states ( @xmath5 ) may be deformed is natural , taking into account the fact that a gas - like @xmath2 state with relatively small number of @xmath0 s could easily be deformed . 5 ) we estimated the critical number of the @xmath0 bosons , @xmath29 , beyond which the system is unbound for the four frameworks , ( gp , dd , s ) , ( gp , pp , s ) , ( hw , pp , s ) and ( hw , pp , d ) . all of the frameworks indicated that the number is roughly @xmath6 , which is not strongly dependent on the frameworks . thus , we concluded that the dilute @xmath2 cluster states could exist in the @xmath3c to @xmath4ca systems with @xmath5 , whose energies vary from threshold up to about 20 mev in the present calculation . the estimate of @xmath6 is of course subject to the validity of our phenomenological approach treating the @xmath0-particles as ideal bosons . we , however , believe , for reasons outlined in the paper , that our estimate for @xmath29 may be correct to within @xmath146 % . in any case the value for @xmath29 , _ i.e. _ the maximum of @xmath0-particles in the condensate state is relatively modest . a very interesting question in this context is whether adding a few neutrons may stabilize the condensate and thus allow for much higher numbers of condensed @xmath0 s . one should remember that @xmath8be is ( slightly ) unbound whereas @xmath147be and @xmath148be are bound . concerning experimental detection of the @xmath0-condensates , the decay scheme of the dilute @xmath2 state is conjectured to proceed mainly via @xmath0 decay . this indicates that such systems may be observed through the following sequential @xmath0 decays : [ dilute @xmath2 state ] @xmath149 [ dilute @xmath150 state]+@xmath0 , [ dilute @xmath150 state ] @xmath149 [ dilute @xmath151 state]+@xmath0 , @xmath152 . therefore , the sequential @xmath0 decay measurement is expected to be one of the promising tools to search for the dilute multi-@xmath0 cluster states , produced via the @xmath0 inelastic reaction , heavy - ion collision reaction , and so on . it is highly hoped to perform such the experiments in near future . we greatly acknowledge helpful discussions with y. funaki , h. horiuchi , k. ikeda , g. rpke , and a. tohsaki . one of the authors ( t.y . ) is very grateful to the theory group of the institut de physique nuclaire , universit paris - sud ( paris xi ) for its hospitality during his sabbatical year . brink , in _ many - body description of nuclear structure and reactions _ , proceedings of the international school of physics `` enrico fermi '' , course 36 , edited by c. bloch ( academic press , new york , 1966 ) . in ref . @xcite , the calculated value is 2.64 fm , which is not taken into account the finite size effect of the @xmath0 particle . with help of the eq . ( [ hw_rms_n ] ) in this paper , the nuclear rms radius @xmath85 is 3.15 fm . . calculated results of the gross - pitaevskii equation with the density - dependent @xmath0-@xmath0 potential in eq . ( [ dd_pot ] ) ; single @xmath0 particle energy @xmath44 and contributions from the kinetic energy @xmath153 , two - range - gaussian term @xmath154 , density - dependent term @xmath155 and coulomb potential @xmath156 in eq . ( [ dd_pot ] ) . the total energy and nuclear rms radius for the n@xmath0 system are denoted as @xmath64 and @xmath85 , respectively . all energies and rms radii are given , respectively , in units of mev and fm . [ cols="^,^,^,^,^,^,^,^,^",options="header " , ] particle potentials @xmath158 ( solid line ) which are obtained by solving the gross - pitaevskii equation with the density - dependent potential ; ( a ) @xmath11 , ( b ) @xmath23 , ( c ) @xmath127 , ( d ) @xmath159 , ( e ) @xmath160 , ( f ) @xmath161 , ( g ) @xmath162 , ( h ) @xmath116 , ( i ) @xmath117 systems . the dashed , dot - dashed and dotted lines demonstrate , respectively , the contribution from the two - range gaussian term , density - dependent term and coulomb potential . ] states measured from each @xmath2 threshold , and ( b ) their nuclear rms radii , which are obtained by solving the gross - pitaevskii equation with the phenomenological @xmath10 and @xmath11 potentials . , title="fig : " ] states measured from each @xmath2 threshold , and ( b ) their nuclear rms radii , which are obtained by solving the gross - pitaevskii equation with the phenomenological @xmath10 and @xmath11 potentials . , title="fig : " ] particle potentials @xmath158 ( solid line ) which are obtained by solving the gross - pitaevskii equation with the phenomenological @xmath10 and @xmath11 potentials ; ( a ) @xmath11 , ( b ) @xmath23 , ( c ) @xmath127 , ( d ) @xmath159 , ( e ) @xmath160 , ( f ) @xmath161 , ( g ) @xmath162 , ( h ) @xmath116 , ( i ) @xmath117 systems . the dashed , dot - dashed and dotted lines demonstrate , respectively , the contribution from the @xmath10 potential , @xmath11 potential and coulomb potential . ] states ( spherical case ) measured from each @xmath2 threshold ( solid line ) , and ( b ) their nuclear rms radii ( solid ) , which are obtained by solving the hill - wheeler equation with the phenomenological @xmath10 and @xmath11 potentials . for comparison , we also give the corresponding values obtained by solving the gross - pitaevskii equation with use of the same potentials ( dotted lines ) . , title="fig : " ] states ( spherical case ) measured from each @xmath2 threshold ( solid line ) , and ( b ) their nuclear rms radii ( solid ) , which are obtained by solving the hill - wheeler equation with the phenomenological @xmath10 and @xmath11 potentials . for comparison , we also give the corresponding values obtained by solving the gross - pitaevskii equation with use of the same potentials ( dotted lines ) . , title="fig : " ] states with the axial deformation measured from each @xmath2 threshold ( solid line ) , and ( b ) their nuclear rms radii ( solid ) and rms distance between @xmath10 particles ( dotted ) , which are obtained by solving the hill - wheeler equation with the phenomenological @xmath10 and @xmath11 potentials . , title="fig : " ] states with the axial deformation measured from each @xmath2 threshold ( solid line ) , and ( b ) their nuclear rms radii ( solid ) and rms distance between @xmath10 particles ( dotted ) , which are obtained by solving the hill - wheeler equation with the phenomenological @xmath10 and @xmath11 potentials . , title="fig : " ]	dilute multi @xmath0 cluster condensed states with spherical and axially deformed shapes are studied with the gross - pitaevskii equation and hill - wheeler equation , where the @xmath0 cluster is treated as a structureless boson . applications to self - conjugate @xmath1 nuclei show that the dilute @xmath2 states of @xmath3c to @xmath4ca with @xmath5 appear in the energy region from threshold up to about 20 mev , and the critical number of @xmath0 bosons that the dilute @xmath2 system can sustain as a self - bound nucleus is estimated roughly to be @xmath6 . we discuss the characteristics of the dilute @xmath2 states with emphasis on the @xmath7 dependence of their energies and rms radii . + + pacs numbers : 21.10.dr , 21.10.gv , 03.75.hh
jenkins j t and mancini f 1987 _ j. appl . mech . _ * 54 * 27	the steady state velocity fluctuations of a movable piston located on the top of a vibrated granular gas are studied by means of molecular dynamics simulations . from the second moment of the distribution , a temperature parameter for the piston is defined and compared with the granular temperature of the gas located just below it . then the two temperature parameters refer to two interacting macroscopic systems . the equipartition of energy valid in usual molecular systems is strongly violated , and the temperature of the piston can be larger or smaller than the one of the gas , depending of the parameters defining the system at the particle level . the simulation results for the ratio of temperatures are in agreement with some theoretical predictions from kinetic theory assuming the validity of a hydrodynamic description in the limit of weak inelasticity of the gas . _ keywords _ : granular matter , kinetic theory of gases and liquids , fluctuations ( theory ) one of the characteristic features exhibited by granular fluids is the violation of the energy equipartition theorem . the granular temperatures of each of the components of a mixture , defined from the average kinetic energies of the particles , are not the same . this was pointed out long ago @xcite , and has attracted a lot of attention in the last decade or so . it is important to stress that the equality of the temperatures of the components of a mixture of molecular fluids is not restricted to equilibrium thermal systems , but also applies to out - of - equilibrium situations , at least to those that can be described by means of usual hydrodynamics as derived from kinetic theory or non - equilibrium statistical mechanics . for the homogeneous cooling state ( hcs ) of a binary granular mixture modeled as an ensemble of smooth inelastic hard spheres , the ratio of temperatures of the two components has been obtained from an approximated solution of the kinetic enskog equations @xcite . this result is in good agreement with molecular dynamics ( md ) simulation data @xcite . the hcs is an idealized state in which the system cools homogeneously and monotonically . in real experiments , to maintain a granular system fluidized , external energy must be continuously supplied , and this is achieved by means of vibrating or rotating walls , by shearing , by shaking , or by other external fields . as a consequence , the systems become inhomogeneous . the non - equipartition has been confirmed by md simulations of simple shear flows @xcite , and of vibrated gases @xcite . there are also some experimental evidences of the coexistence of different component temperatures in strongly vibrated granular mixtures @xcite . homogeneously driven granular fluids submitted to stochastic forces have also been considered . the lack of energy equipartition in these systems has been analyzed @xcite , but the relationship between this kind of ideal driving and actual experiments is uncertain . a peculiarity of the above studies is that they are concerned with quantities going beyond the macroscopic description , in the sense that the relevant macroscopic local property is the global temperature ( or energy density ) of the mixture , and not that of each of the components . moreover , the latter are rather difficult to measure experimentally . here , a possible experimental set - up in which the breakdown of energy equipartition is macroscopically observed is described and studied by means of md and kinetic theory . the experiment is specifically designed to be reproducible in the laboratory , and might have some practical advantages over previous related experiments @xcite , although it must be stressed that the study reported here is restricted to low density gases and that no friction between the particles , of the particles with the piston , or of the latter with the walls of the container is taken into account . consider a granular system composed of @xmath0 inelastic hard spheres ( @xmath1 ) or disks ( @xmath2 ) of mass @xmath3 and diameter @xmath4 . there is an external gravitational field acting on the system , so that each particle is submitted to a force @xmath5 , where @xmath6 is a positive constant and @xmath7 the unit vector in the direction of the positive @xmath8 axis . the system is kept fluidized by vibrating the lower wall of the container in a sawtooth way , with very high frequency , negligible amplitude , and velocity @xmath9 @xcite . the latter will be always taken large enough as to keep the density of the system small everywhere . on the top of the system there is a movable lid or piston of mass @xmath10 , as illustrated in fig . [ fig1 ] . the piston can only move in the @xmath8-direction , remaining always perpendicular to it . it is assumed that there is no friction between the piston and the lateral walls of the vessel containing the granular gas . collisions between particles are inelastic and characterized by a constant coefficient of normal restitution @xmath11 , so that when a pair of particles @xmath12 and @xmath13 collides , their relative velocity @xmath14 changes instantaneously according with @xmath15 while the total momentum is conserved . in the above expression , @xmath16 is a unit vector directed from the center of particle @xmath13 to the center of particle @xmath12 at contact . collisions of particles with the piston are also modeled as smooth and inelastic , being @xmath17 the coefficient of restitution in this case . taking into account that the piston can only move in the @xmath8 direction , the change in the relative velocity in a collision is restricted to the @xmath8-component , and it is given by @xmath18 where @xmath19 is the velocity of the piston . this relation and the total momentum conservation define the collisions with the piston . next , some results from md simulations of two - dimensional systems will be presented . to avoid undesired effects , periodic boundary conditions are used in the direction perpendicular to the external field . it is observed that , at least when the width @xmath20 of the system is not too large , a steady state with no macroscopic flow of mass , and gradients in the fluid only in the direction of the external field , i.e. the @xmath8 axis , is reached after some transitory period . this state is independent from the value of @xmath20 , as long as it is kept small enough . the shape of the hydrodynamic profiles in the bulk will be analyzed elsewhere @xcite , while here the emphasis will be put on the relationship between the temperature parameters of the piston , @xmath21 , and of the gas in its vicinity , @xmath22 . both temperatures are defined from the second moment of the respective velocity distributions , namely @xmath23 , @xmath24 , with the angular brackets denoting average . in the results to be reported in the following , units defined by @xmath25 , @xmath26 , and @xmath27 will be used . the measured probability distribution for the velocity of the piston is very well fitted by a gaussian in all the cases considered . a typical example , corresponding to @xmath28 , @xmath29 , @xmath30 , @xmath31 , @xmath32 , and @xmath33 , is given in fig . [ fig2 ] . for this case , the value @xmath34 is found . it is seen that the gaussian fit is accurate at least up to values of the probability density of the order of @xmath35 . the measurement of @xmath22 is somewhat more problematic , since the ` top region ' of the granular gas next to the piston has to be identified . here , the layer of the gas defined by @xmath36 , where @xmath37 is the dispersion of the values of the position of the piston @xmath38 , has been used . then , the average over the velocities to compute @xmath22 has been carried out over the particles inside this region at each time . this criterium may appear as rather arbitrary , but it has been checked that other sensible choices , such as extrapolating the bulk temperature profile to @xmath39 , lead to values of the temperature with discrepancies of the order of @xmath40 . in this way , it is found that @xmath41 , implying @xmath42 . this represents a strong violation of the energy equipartition , specially taking into account that the collisions between the grains and the piston have been modeled as elastic ( @xmath43 ) . moreover , it must be stressed that the temperature parameter of the piston is larger than the temperature of the gas transmitting to it the effects of the lower vibrating wall . it can be wondered whether this effect is related with some strong anisotropy of the velocity distribution of the gas . let us define @xmath44 and @xmath45 , so that @xmath46 . then , for the particular case we are discussing , it is found that @xmath47 . therefore , although there is some anisotropy as usual in vibrated granular systems , it is very weak as compared with the excess value of @xmath48 . more precisely , it is @xmath49 . the same trend is observed in all the other cases investigated , being the above the one in which the anisotropy of the second moment of the velocity tensor is stronger . in fig . [ fig3 ] , the temperature ratio @xmath48 is plotted as a function of the coefficient of restitution of the gas @xmath11 for three different values of the pair ( @xmath10 , @xmath17 ) , illustrating the influence of each of these parameters on the temperature ratio . quite interestingly , the violation of equipartition ( deviation of the temperature ratio from unity ) is not a monotonic function of @xmath11 , but the temperature ratio exhibits a minimum in the small inelasticity region . the result with @xmath50 , @xmath30 , has been obtained with a non - vibrated system at thermal equilibrium . as expected , equipartition is recovered in the limit @xmath51 . the reason to restrict the analysis to the inelasticity region @xmath52 is that , for smaller values , even the hydrodynamic description of the bulk of the gas , as provided by the navier - stokes equations , breaks down @xcite . an expression for the temperature ratio can be derived as follows . given the low density of the gas , a pair of coupled boltzmann and boltzmann - lorentz kinetic equations are used as the starting point to describe both the gas and the piston . the subsequent calculations are based on the following two assumptions : ( a ) the macroscopic state of the piston can be specified in terms of the hydrodynamic fields of the gas in its vicinity and its position probability distribution , and ( b ) the associated normal solution of the kinetic equations can be generated by the chapman - enskog procedure . with these hypothesis , it is obtained that the zeroth order in the gradients cooling rates of the piston , @xmath53 and of the top region of the gas , @xmath54 , must be the same @xcite . these rates are functionals of the one - particle distributions for the top region of the fluid and the piston . given that these distributions are not known , maxwellians can be used as an estimation . therefore , no anisotropy effects are introduced in the theory . a generalization of the derivation in @xcite , with the peculiarity that now all the the collision vectors have the same direction , leads to @xcite @xmath55 @xmath56 where @xmath57 is the gas density in the top region , @xmath58 , @xmath59 , and @xmath60 note that all the dependence on the mass ratio and the inelasticity @xmath17 occurs through @xmath61 . requiring the cooling rates to be the same , an equation determining @xmath62 is obtained , @xmath63 the elastic collisions limit ( @xmath64 ) is given by @xmath65 , that is the equipartition result . if the particle collisions are elastic but not those with the piston ( @xmath66 ) , it is @xmath67 $ ] . this result is the same as the one derived in ref . @xcite for an inelastic intruder in an equilibrium elastic gas . the prediction made by eq . ( [ 6 ] ) is compared with some md simulation results in fig . [ fig4 ] , where @xmath62 is plotted as a function of @xmath61 . in the simulations , the values of the latter quantity have been modified by changing both @xmath10 ( in the range @xmath68-@xmath69 ) and @xmath17 ( between @xmath70 and @xmath71 ) . the other parameters ( @xmath20 , @xmath72 , and @xmath0 ) are kept the same as in fig . . the symbols are the simulation results while the lines are from eq . ( [ 6 ] ) , as indicated . data for two values of @xmath11 , @xmath73 and @xmath74 , are reported . for the most elastic gas , the agreement is quite good , for all the considered values of the other parameters . nevertheless , when the inelasticity of the gas is slightly increased , strong deviations from the theoretical prediction show up . these discrepancies are not only quantitative , but also qualitative . the circled symbols on the left lower corner of the figure correspond all them to @xmath31 , differing in the value of @xmath17 . it is observed that when @xmath17 ( and @xmath61 ) increases , @xmath62 decreases , in contradiction with the solution of eq . ( [ 6 ] ) . to identify the cause of the failure of the theory , the shape of the marginal distribution of the component of the velocity of the grains perpendicular to the piston in the top region , @xmath75 , has been measured . while for @xmath76 it is always well fitted by a gaussian within the numerical error of the simulations , for @xmath77 , exponential tails are clearly identified for positive velocities already in the thermal range , as it is seen in fig . [ fig5 ] . these tails are relevant for the lowest velocity moments of the gas . on the other hand , it was already mentioned above ( see fig . ) that the velocity distribution of the piston remains gaussian . a theoretical analysis of the existence of this exponential tail probably requires going beyond the chapman - enskog method , the restriction to normal solutions of the boltzmann equation and , therefore , of the hydrodynamic description . the above results seem to provide additional evidence of the difficulties when trying to give a macroscopic meaning to the granular temperature , at least as defined in kinetic theory and the usual hydrodynamic theories . the knowledge of the hydrodynamic profiles of the confined gas and the characteristics of the interactions between the gas particles and the piston is not enough to determine the `` temperature '' of the latter . more detailed information about the velocity distribution of the gas in its vicinity is required . moreover , it is surprising that sometimes the `` temperature '' of the body to which energy is supplied ( the piston ) be larger than the temperature of the system supplying the energy ( the neighbor gas ) . the results also show that this effect is not associated to the anisotropy of the gas . finally , let us stress the differences of this system as compared with a gas mixture , even in the tracer ( one - intruder ) limit . here , two macroscopic systems spatially differentiated are interacting at contact , their respective states being strongly interdependent . as a consequence , it seems clear that both temperature parameters have a macroscopic meaning . this research was supported by the ministerio de educacin y ciencia ( spain ) through grant no . fis2005 - 01398 ( partially financed by feder funds ) .
ising spin models are probably the simplest interacting many particle systems to which many equilibrium and non - equilibrium phenomena found in nature can be mapped . in many cases the analysis of these models exhibit the basic features and consequently lead to a first understanding of the reality . therefore numerous investigations employing ising models exist in literature and are the subject of research activities . in addition to the questions in physics various problems arising in other fields ( biology , social science , information theory , etc . ) are treated on the basis of ising models . ising spin systems do not exhibit an intrinsic dynamics . thus the kinetics of these models is exclusively caused by couplings to bath systems of the environment . forty years ago glauber @xcite suggested a phenomenological master equation for a one - dimensional chain of ising spins with nearest neighbor interactions . generalized to other spin arrangements and to other types of interactions ( for a review see @xcite ) the approach of glauber has been nearly exclusively used to describe the dynamics of the ising spin systems . note that the above comment also applies to the numerous monte carlo simulations performed on various models in the last decades @xcite . the transition rates of these phenomenological equations are required to satisfy detailed balance which guaranties relaxation to equilibrium for non driven systems . but this requirement is not sufficient to fix the rates completely . therefore ad hoc assumptions have to be employed which lead to the claim @xcite that kinetic ising models should not be regarded as faithful representations of processes occurring in real systems . it is microscopic quantum - statistical investigations which can remove the arbitrarily of the phenomenological equations . the tools for such a purpose are known since nearly half a century ago @xcite and many general approaches exist . the application of these general theories in ising models , however , are rare . a first approach of heims @xcite uses a very special and thus artificial spin - bath interaction . further approaches , both based on special realizations of the heat bath , are presented by martin @xcite and recently by park et al.@xcite . all these microscopic approaches find transition rates different from @xcite . a recent investigation for the linear response of a general ising model @xcite based on the mori theory results in the same conclusion . it is the aim of the present letter to calculate the transition rates under general aspects and to work out the generic behavior for the general non - linear case . thus neither specific ising models nor specific realizations of the baths are presumed . the ising spins may even interact with more than one bath to include problems like those treated in @xcite . calculations are performed on the bases of the robertson theory @xcite , although equivalent result @xcite are obtained by employing @xcite or the nakajima - zwanzig approach @xcite . with rare exceptions like the original problem of glauber @xcite , analytical solutions of the master equations are not known and therefore approximations have to be employed . the typical procedure involves the elimination of fast degrees of freedom in the master equation . this leads to a reduced description and equations of motion result for the slow or relevant variables . the latter equations , in many cases , can directly be found by the robertson theory . thus both , the master equation and the equation of motion of such reduced descriptions can simultaneously be derived which justifies the above choice of the method . a system of @xmath0 quantum spins @xmath1 with @xmath2 is considered in the presence of time dependent external fields @xmath3 . the spins interact via an arbitrary ising spin - spin interaction @xmath4 and are described by the spin hamiltonian @xmath5 where @xmath6 is presumed . both quantum spin @xmath7 operators @xmath8 with @xmath9 and ising spins @xmath10 are used simultaneously . the assembly of spins interact via spin bath interactions @xmath11 with a system consisting of @xmath12 adiabatically isolated baths @xmath13 . has no effect and that the coupling of @xcite or similar two spin couplings will lead to spatial correlations via @xmath14 . such a behavior may be possible but is certainly not the generic case . ] thus the total hamiltonian of the system is described by @xmath15 during the dynamic evolution of the spin system thermal equilibrium is presumed for each individual bath system @xmath16 leading to the statistical operator @xmath17 these requirements on the bath systems can be realized by sufficient large bath heat capacities and by faster relaxation of the bath system . thus @xmath18 where @xmath19 and @xmath20 are the characteristic relaxation times of the bath and the spin system respectively . in general all the temperatures @xmath21 are different . due to couplings to additional reservoirs the @xmath22 may even be time dependent . thus the @xmath23 - together with @xmath24 - are preset quantities . both the @xmath23 and the @xmath24 are presumed to have negligible variations on the fast time scale @xmath25 . ) is an approximation which requires this presumption for the @xmath24 . ] with these requirements there is no need for an explicit specification of the bath hamiltonian @xmath26 and the bath operators @xmath27 . as shown below , it is just the absorptive part @xmath28 of the dynamic bath susceptibility @xmath29\rangle_b\ ; e^ { i\omega t } \mathrm{d}t \quad,\end{aligned}\ ] ] which enters the calculation for values @xmath30 . expectation values with respect to @xmath31 are represented by @xmath32 and @xmath33 denotes the liouvillian defined by @xmath34 $ ] . it is assumed that @xmath35 holds , which can always be achieved by a renormalization of the terms of the hamiltonian . note that @xmath36 implies @xmath37 and the first term of the low frequency expansion of @xmath38 can be used . apart from interesting exceptions ( compare @xcite ) this yields in the generic case @xmath39 where the constant of proportionality may depend on @xmath40 . let @xmath41 be a fixed but arbitrary set of variables which are functions of the ising spins @xmath42 . this set is assumed to give a sufficient dynamical description of some physical question and will be called observation level according to @xcite . without further specification of the observation level at this stage the general approach of robertson @xcite is applied employing the usual perturbation treatment up to the second order in @xmath43 and the standard markovan approximation @xcite . as all variables @xmath44 commute with @xmath45 this leads to @xmath46 \rangle_t\quad , \quad u(t , t-\tau)= \exp ( -i\tau[l^s(t)+l^b ] ) { \end{equation}}where the liouvillians $ l^s $ and $ l^b $ are associated to $ { \cal h}^s $ and to $ { \cal h}^b $ , respectively . the time dependent expectation values $ \langle \ldots \rangle_t $ are performed with the generalized canonical statistical operator { \begin{equation}\label}{54 } r_t = r^s_t\ : r^b \quad \mathrm{with}\quad r^s_t={\exp ( v ) } / { { \rm tr_s}\;\exp(v ) } \quad \textrm{and}\quad v=\sum_k \lambda^k_t a^k \quad\ ] ] where the time dependent lagrange parameters @xmath47 are implicitly determined by @xmath48 note that the system of eqs.([50 ] ) and ( [ 11 ] ) represent a closed set of nonlinear differential eqs . for the expectation values @xmath49 and the lagrange parameters @xmath50 . thus together with the initial values for the @xmath51 ( or with the initial values for @xmath50 ) the time dependence of all quantities are completely determined by this set of equations . analog to @xcite eq.([50 ] ) can be brought in a more convenient form . using again @xmath52 and @xmath53 this leads to @xmath54}\big\rangle_t \ : \quad \\\label{54}&\quad = \sum_{\alpha i}\big \langle \gamma_{\alpha i}\big\ { \tanh ( \beta_\alpha h_i+\beta_\alpha x_i ) - \tanh v_{[i]}\big \ } a^k_{[i]}\big \rangle_t\end{aligned}\ ] ] with @xmath55 for all functions @xmath56 of the @xmath0 ising spins @xmath57 } $ ] is defined as @xmath58}=\frac{1}{2}{\rm tr}_i \ ; s_i\ ; f ( s_1,\ldots , s_n)\ ] ] which enter in the calculation from @xmath59=\mp 2 s_i^\pm f_{[i]}$ ] . the operators of the internal field at sites @xmath60 are @xmath61}=\sum_j j_{ij}s_j\quad.\ ] ] in the step leading to eq.([54 ] ) the relation @xcite @xmath62 } ) \,\rangle_t\ ] ] was used which holds for any operator @xmath63 of the ising spins not involving the spin @xmath64 . eqs.([53 ] ) and ( [ 54 ] ) represent in a compact form the most general results of this work . these results are valid for any observation level which has to be specified for concrete physical problems . as already pointed out , all slow variables of the specific problem have been included . one can also see that via eq.([54 ] ) a second requirement on the set of variables results . being , in principle , arbitrary the initial value of @xmath65 is assumed to have the special form of eq.([54 ] ) . by an extension of the observation level , however , this second requirement can always be satisfied for any given initial state . here two special observation levels will be discussed . first , the observation level is analyzed which leads to the master equations and contains _ all _ functions @xmath66 of the n ising spins . obviously all of the above requirements are satisfied . the second observation level is spanned by all ising spins @xmath67 and by the interaction @xmath68 . this choice implies that all thermal equilibrium states of the spin system are possible initial states . consequently this observation level leads to a dynamical description of a standard thermodynamic system and is therefore denoted as standard system in the following . let the @xmath69 be the common eigenvectors of the @xmath64 where @xmath70 denotes the configuration of the spin system with the possible values @xmath71 . then the observation level is spanned by all the @xmath72 projectors @xmath73 . introducing the notation @xmath74 , the relation @xmath75}=\sigma_i/2\ , \big(\|\bm{\sigma}\rangle\langle\bm{\sigma}\|\,+\,\,\|\bm{\sigma}^{(i)}\rangle\langle\bm{\sigma}^{(i)}\| \big ) $ ] holds . eq.([53 ] ) immediately leads to the master equation @xmath76 for the occupation probabilities @xmath77 . the transition rates @xmath78 are given by @xmath79}\ ] ] with @xmath80 _ the rates @xmath81 _ are consistent with the results of @xcite but _ disagree with the usual ad hoc assumption of phenomenological kinetic ising models _ @xcite which uses constant values of coefficient @xmath82 independent of the configuration @xmath83 . thus for a faithful representation of processes occurring in real systems _ the full @xmath83 dependence of the @xmath82 has in general be used . _ this observation level will be characterized by @xmath84 which leads to @xmath85}= \lambda_i + \lambda_{ex } x_i $ ] . introducing the internal local - field probability functions @xmath86 and employing again the relation ( [ 54 ] ) the expectation values of the magnetizations @xmath87 and the exchange energy @xmath88 can be written as ( compare @xcite ) @xmath89 and the eqs.([53 ] ) leads to @xmath90\label{66}\\ & \frac{d \,}{d t}\,u^{ex } \,= \sum_{\alpha i}\int\textrm{d}x \,x\ , w_i(x ) \frac{\chi''_{\alpha i } ( 2 h_i + 2 x ) } { 2 } \big [ \frac{\tanh ( \lambda_i + \lambda_{ex}x ) } { \tanh ( \beta_\alpha h_i + \beta_\alpha x ) } -1\big]\label{67}\:.\end{aligned}\ ] ] _ the set of nonlinear equations _ ( [ 65]-[67 ] ) for the dynamic variables @xmath91 and @xmath92 together with the associated time dependent lagrange parameters @xmath93 and @xmath94 _ is closed and is expected to give an adequate description for many physical systems and situations . _ note that the internal local - field distribution functions @xmath95 govern the set of nonlinear equations of motion . indeed all the terms of these set of equations can explicitly be calculated from the knowledge of the @xmath96 . these findings generalize the previous results , that the @xmath96 determine the statics @xcite and the dynamical linear response @xcite of an arbitrary ising model . in that previous work the distributions @xmath97 have explicitly been calculated for various specific physical models including both ferromagnetic and disordered systems . consequently for all those models the nonlinear equation of motion can easily be obtained from the eqs.([65]-[67 ] ) . the latter procedure will be illustrated for infinite range spin glass model of sherrington and kirkpatrick @xcite . in the sk model the bonds @xmath98 are independent random variables with zero means and standard deviations @xmath99 . this scaling fixes the spin glass temperature to @xmath100 . employing the modified tap approach@xcite @xmath101 holds @xcite with @xmath102 where @xmath103 is determined from @xmath104 with the eqs.([80]-[83 ] ) the equations of motion ( [ 65]-[67 ] ) are explicit provided the system of baths is specified . for the case of one singular bath and identical spin bath couplings the sums over @xmath105 drop out in eqs.([66]-[67])and only one bath temperature @xmath106 remains . the resulting equations are expected to describe the complete non linear dynamics of the sk spin glass including temperature quenches , memory and ageing effects @xcite . a complete discussion of these interesting effects is beyond the scope of this work and is a subject of further research . in the following , i focus on the discussion of the special situations when the exchange system is in equilibrium with the bath , implying @xmath107 . this situation can physically be realized by allowing only changes of the fields @xmath108 . setting @xmath109 eq.([66 ] ) yields glauber type equations of motion @xmath110 \quad \quad , \quad \quad h_i^{eff}=h_i + \sigma_j j_{ij}m_j - m_i\beta \delta\ ] ] with the relaxation rates @xmath111 where @xmath112 has to be determined from eq.([81]-[83 ] ) ( with @xmath113 replaced by @xmath114 ) . on a phenomenological basis - with @xmath115 replaced by a constant - equations of motions of this kind have been used since the early days in spin glass research @xcite . moreover the author himself had used such equations to find numerically the solutions of the static tap equations @xcite . note that eq.([91 ] ) implies relaxation to the static tap equations independent of the values of @xmath116 . thus the phenomenological equations can be used instead of the microscopic results to calculate static properties and these results justify the procedure used in @xcite to calculate the solutions of the tap equations . for all dynamic quantities , however , the full form of @xmath116 has to be used . first examples are the dynamic susceptibility and the relaxation function as presented already in @xcite . note that a result of @xcite , the frequency - dependent shift of the cusp temperature of the real part of the susceptibility , is basically a consequence of @xmath116 . in this context the general result @xcite should be recalled that the linear approximation of the robertson theory leads the mori theory . consequently all the linear response theory results of @xcite can also be obtained from the of eqs.([90 ] ) and ( [ 91 ] ) by linearizing around the thermal equilibrium which can easily be checked . the consequences of the results of this letter are further illustrated on simple physical situations . first one singular spin coupled to a bath with temperature @xmath114 is considered . this implies no spin spin interaction at all . both the master equation and the standard system approach of this letter leads to the equation of motion @xmath117 for the magnetization @xmath118 with @xmath119 . for a static field the magnetization @xmath118 relaxes to the equilibrium value @xmath120 with the well known ( see e.g. @xcite ) longitudinal relaxation rate @xmath121 . note the latter result is not found if instead of eq.([g ] ) the usual ad hoc assumptions of the kinetic ising model are used . next the original system of glauber , an infinite chain of ising spins with next neighbor ferromagnetic interactions is considered which is coupled to one bath with a temperature @xmath122 . the transition rates ( [ g ] ) satisfy the detailed balance relation and therefore relaxation to the thermal equilibrium results within the master equation approach . the same consequences apply for the standard system approach . indeed the equilibrium values of the lagrange parameter @xmath123 and @xmath124 describe a fixed point of the eqs.([66 ] ) and ( [ 67 ] ) . it is obvious that nether the transition rates ( [ g ] ) nor the details of distribution functions @xmath125 enter in thermal expectation values . note that these arguments apply to all undriven ising system which are coupled to one singular bath . thus all this systems relax to thermal equilibrium taking in account that the fixed points are always stable . this can easily be checked by a linear stability analysis . for situations , however , where two or more bathes with different temperatures @xmath126 are present the values of stationary solutions depend on the transition rates or on the details of the distribution functions @xmath96 . thus for example the non equilibrium stationary states of schmittmann and schmser @xcite can not be accepted as a realistic solution for their interesting model of a one dimensional ising chain coupled to two bath systems as the analysis is based on the phenomenological rates and not on the physical rates . finally it is pointed out that the results of the present work will always effect the details of the dynamic evolution apart form situations where the hight temperature approximation can be used . in general more complicated expressions are found compared to the phenomenological results . this even applies to the original model of glauber . for this kinetic ising model the equations of motion for the time correlation function partly separate which enables the explicit calculation of the complete dynamics . such a simplifying separation does not occur for the microscopical master equation . thus the solution is not known and is hard to find as already pointed out in @xcite . interesting discussions with b. drossel , w. just , g. sauermann and f. schmser are acknowledged . 20 glauber r. j. , _ j math phys _ , * 4 * ( 1963 ) 294 . kawasaki k. , _ phase transitions and critical phenomina , eds . : domb c. and green m. s. _ p 443 ( london , new york : academic press ) 1972 . binder k. , _ monte carlo methods in statistical physics _ , p 1 ( berlin heidelberg : springer ) 1986 . bloch f. , _ phys . _ , * 105 * ( 1957 ) 1206 ; redfield a. _ ibm j. res . _ , * 1(1957 ) 17 . heims s. p. , _ phys . rev . _ , 138 * ( 1965 ) a587 . martin p. a. , j. stat . phys . * 16 * ( 1977 ) 149 . park k. novotny m. a. , comput . * 147 * ( 2002 ) 737 ; park k. , novotny m. a. rikvold p.a . _ e , * 66 * ( 2002 ) 056101 . plefka t. , _ j. phys . a : math . gen._,35 ( 2002 ) 8691 . schmser f. schmittmann b.,_j . phys . a : math . gen._,35(2002 ) 2569 ; schmittmann b. schmser f. , _ phys _ e , * 66 * ( 2002 ) 046130 . robertson b. , _ phys . _ , * 44 * ( 1966 ) 151 . fick e. sauermann g. , _ the quantum statistics of dynamic processes _ ( berlin heidelberg new york : springer - verlag ) 1990 . just w. private communication . elmer f.j . fick e. , _ physica a _ , * 117 * ( 1983 ) 232 ; plefka t. , _ physica a _ , * 162 * ( 1990 ) 477 ; plefka t. , _ z. phys.b _ , * 90 * ( 1993 ) 447 . thomsen m. , thorpe m. f. , choy t. c. sherrington d. , _ phys.rev._ , * b 30 * ( 1984 ) 250 , ; thomsen m. , thorpe m. f. , choy t. c. , sherrington d. sommers h .- j . , _ phys.rev._ , * b 33 * ( 1986 ) 1931 . sherrington d. kirkpatrick s. _ phys . _ , * 32 * ( 1975 ) 1972 . mezard m. , parisi g. virasoro m. a. , _ spin glass theory and beyond _ ( singapore : world scientific ) 1987 and references therein . yong a. p. , _ spin glasses and random fields _ ( singapore : world scientific ) 1997 . plefka t. , _ europhys . _ , * 58 * ( 2002 ) 892 . kinzel w. fischer k. h. , _ solid state commun . _ , * 23 * ( 1977 ) 687 . plefka t. , _ phys . _ b , * 65 * ( 2002 ) 224206 .	based on the robertson theory the nonlinear dynamics of general ising systems coupled microscopically to bath systems is investigated leading to two complimentary approaches . within the master equation approach microscopically founded transition rates are presented which essentially differ from the usual phenological rates . the second approach leads to coupled equations of motion for the local magnetizations and the exchange energy . simple examples are discussed and the general results are applied to the sherrington - kirkpatrick spin glass model .
the paper is concerned with * an existence theorem for the asymptotics of evaluations of arbitrary spin networks ( using the theory of @xmath0-functions ) , * a rationality property of the generating series of all evaluations with a fixed underlying graph ( using the combinatorics of the chromatic evaluation of a spin network ) , * a complete study of the asymptotics of the so - called @xmath1-symbols in all cases : euclidean , plane or minkowskian ( using the theory of borel transform ) , * the assignement of two numerical invariants to spin networks and cubic graphs , the spectral radius and a number field ( see definition [ def.radius ] ) , * an explicit illustration of our results for the regular and some flat @xmath1-symbols ( using the constructive wz method ) . a ( classical ) _ spin network _ @xmath2 consists of a _ cubic ribbon graph _ @xmath3 ( i.e. , an abstract trivalent graph with a cyclic ordering of the edges at each vertex ) and a coloring @xmath4 of its set of edges by natural numbers . according to penrose , spin networks correspond to a diagrammatic description of tensors of representations of @xmath5 , and their evaluation is a contraction of the above tensors . spin networks originated in racah s work in atomic spectroscopy in the late forties @xcite and in the work of wigner in the fifties @xcite and their evaluations ( exact or asymptotic ) is a useful and interesting topic studied by ponzano - regge , biedenharn - louck and many others ; see @xcite . in the past three decades , spin networks have been used in relation to classical and quantum gravity and angular momentum in 3-dimensions ; see @xcite . on the mathematical side , @xmath6-deformations of spin networks ( so called quantum spin networks ) appeared in the eighties in the work of kirillov - reshetikhin @xcite . _ quantum spin networks _ are _ knotted framed trivalent graphs _ embedded in 3-space with a cyclic ordering of the edges near every vertex , and their evaluations are rational functions of a variable @xmath6 . the quantum @xmath7 and @xmath1-symbols are the building blocks for topological invariants of closed 3-manifolds in the work of turaev - viro @xcite . quantum spin networks are closely related to a famous invariant of knotted 3-dimensional objects , the celebrated _ jones polynomial _ , @xcite . a thorough discussion of quantum spin networks and their relation to the jones polynomial and the _ kauffman bracket _ is given in @xcite and @xcite . aside from the above applications , spin networks and their evaluations have an intrinsic interest , they are formulated in terms of elementary combinatorial data and offer many challenges even for relatively simple networks such as the cube . some examples of spin networks that will be discussed in the paper are shown in figure [ fig.exspin ] . @xmath8 spin networks with underlying graph @xmath9 or a tetrahedron are usually called _ @xmath7-symbols _ and _ @xmath1-symbols _ respectively and they form the building blocks for the evaluation of any spin network ; see for example section [ sub.recoupling ] or @xcite . we now discuss the evaluation of a spin network , introduced by penrose in @xcite . \(a ) we say a spin network is _ admissible _ when the sum of the three labels @xmath10 around every vertex is even and @xmath10 satisfy the triangle inequalities : @xmath11 . ( b ) the evaluation @xmath12 of a spin network @xmath2 is defined to be zero if it is not admissible . an admissible spin network is evaluated by the following algorithm . * use the cyclic ordering to thicken the vertices into disks and the edges into untwisted bands . * replace the vertices and edges by the linear combinations of arcs as shown in figure [ fig.evaluation ] . * finally the resulting linear combination of closed loops is evaluated by assigning the value @xmath13 to a term containing @xmath14 loops . @xmath15 for example , in figure [ fig.exspin ] the @xmath9 network and the cube are admissible , the @xmath16-sided drum and the @xmath17-symbol are not . note that the admissibility condition is equivalent to saying that the strands can be connected at each vertex as in figure [ fig.evaluation ] . note also that cubic ribbon graphs @xmath3 are allowed to have multiple edges , loops and may be disconnected with components that contain no vertices . in addition , @xmath3 may be and non - planar ( contrary to the requirement of many authors @xcite ) , as long as one fixes a cyclic ordering of the edges at each vertex . the latter condition is implicit in @xcite . it turns out that a different cyclic ordering on a vertex of a spin network effects its evaluation by a sign ; see lemma [ lem.evaluation ] below . it is easy to see that if @xmath2 is an admissible spin network and @xmath14 is a natural number , then @xmath18 is admissible as well . a fundamental problem is to study the asymptotic behavior of the sequence of evaluations @xmath19 when @xmath14 is large . this problem actually consists of three parts . fix an admissible spin network @xmath2 . * prove the existence of an asymptotic expansion of the sequence @xmath20 when @xmath14 is large . * compute the asymptotic expansion of @xmath20 . * identify the terms in the asymptotic expansion with geometric invariants of the spin network . although any spin network can be evaluated in terms of @xmath7 and @xmath1-symbols , a positive solution to parts ( a ) , ( b ) or ( c ) of problem [ prob.key ] for @xmath7 and @xmath1-symbols does _ not _ imply a solution to for arbitrary spin networks . problem [ prob.key ] is motivated by the belief that the quantum mechanics of particles with large spin will approximate the classical theory . there is a large physics literature for part ( b ) ; see for example @xcite . on the other hand , the literature for part ( a ) is relatively new and short , and to the best of our knowledge concerns only @xmath7 and @xmath1-symbols with certain labellings . roberts used geometric quantization techniques to prove the leading asymptotic behavior of @xmath1-symbols in the so - called euclidean case ; see @xcite . in order to get an exponentially ( and not factorially ) growing asymptotic expansion we will rescale the evaluations of spin networks by factorials of linear forms on the labels of their edges as follows . we define the _ standard normalization _ of a spin network evaluation to be @xmath21 where @xmath22 is defined to be the product @xmath23 where @xmath24 are the colors of the edges emanating from vertex @xmath25 , and @xmath26 is the set of vertices of @xmath3 . there are many useful properties of this evaluation : * @xmath27 is still an integer ( see theorem [ thm.gf ] ) , * the sequence @xmath28 is exponentially bounded ( see theorem [ thm.existence ] ) , * the spin generating series is a rational function ( see theorem [ thm.gf ] below ) . our first result is a complete solution of part ( a ) of problem [ prob.key ] for all spin networks , using the theory of @xmath0-functions . to state it , we need to recall the notion of a sequence of nilsson type @xcite . we say that a sequence @xmath29 is of _ nilsson type _ if it has an _ asymptotic expansion _ of the form @xmath30 where * the summation is over a finite set , * the _ growth rates _ @xmath31 are algebraic numbers of equal magnitude , * the _ exponents _ @xmath32 are rational and the _ nilpotency exponents _ @xmath33 are natural numbers , * the _ stokes constants _ @xmath34 are complex numbers , * the formal power series @xmath35 $ ] are gevrey-1 ( i.e. , the coefficient of @xmath36 is bounded by @xmath37 for some @xmath38 ) , * @xmath39 is a _ number field _ generated by the coefficients of @xmath40 for all @xmath41 . the above definition uses the notion of an asymptotic expansion in the sense of poincare , given in ( * ? ? ? * sec.1 ) . in the special case where there is only one growth rate , the notation @xmath42 means that for every @xmath43 we have @xmath44 in section [ sec.hyper ] we will prove the following . for every spin network @xmath45 , the sequence @xmath46 is of nilsson type . theorem [ thm.existence ] follows from the theory of @xmath0-functions by combining theorems [ thm.g ] and [ thm.gtaylor ] below . @xmath0-functions were introduced by siegel ( @xcite ) in relation to transcendence problems in number theory . many of their arithmetic and algebraic properties were established by andr in @xcite . @xmath0-functions appear naturally in geometry ( as variations of mixed hodge structures ) , in arithmetic and most recently in enumerative combinatorics . for a detailed discussion , see @xcite and references therein . let us recall the notion of a @xmath0-function . we say that series @xmath47 is a _ @xmath0-function _ if * the coefficients @xmath48 are algebraic numbers , * there exists a constant @xmath38 so that for every @xmath49 the absolute value of every conjugate of @xmath48 is less than or equal to @xmath50 , * the common denominator of @xmath51 is less than or equal to @xmath50 , * @xmath52 is holonomic , i.e. , it satisfies a linear differential equation with coefficients polynomials in @xmath53 . the following theorem is proven in section [ sec.hyper ] . the proof is elementary given the results of @xcite and the fact that an evaluation @xmath54 can be written as a multi - dimensional combinatorial sum of a balanced hypergeometric term . the generating function @xmath55 of an arbitrary spin network @xmath2 is a @xmath0-function . the next result follows by a combination of results of andr , katz and chudnovsky - chudnovsky ; see @xcite and also ( * ? ? ? * sec.2 ) . if @xmath47 is a @xmath0-function , then @xmath29 is a sequence of nilsson type . every @xmath0-function has two basic invariants : an analytic and an arithmetic one : * the radius of convergence at @xmath56 , which equals to @xmath57 where @xmath31 is a growth rate of @xmath29 as in definition [ def.nilsson ] , * a number field @xmath39 corresponding to @xmath29 as in definition [ def.nilsson ] . theorem [ thm.g ] motivates the following invariants of a spin network : its spectral radius and its number field . let @xmath2 be a spin network . * the _ spectral radius _ @xmath58 is defined to be the inverse of the radius of convergence of the generating function @xmath59 at @xmath56 . * the _ number field _ @xmath60 is defined to be the number field associated to the @xmath0-function . * the _ spectral radius _ @xmath61 and the number field @xmath62 of a cubic ribbon graph is defined to be @xmath63 and @xmath64 respectively , where @xmath65 denotes the admissible coloring of all edges of @xmath3 by @xmath65 . the proof of theorem [ thm.existence ] is not constructive . we know of several different methods for computing the asymptotic expansions of spin network evaluations . we will list them here briefly and illustrate them in the following sections . * the method of _ borel transform _ for one - dimensional sums of balanced hypergeometric terms . this method has been developed for one - dimensional sums in @xcite ) and gives exact formulas for the complete asymptotic expansion of such sums . the method ought to extend to multi - dimensional sums , but this still needs to be developed . since @xmath1-symbols are given by one - dimensional sums of a balanced hypergeometric term ( in all cases , euclidean , plane or minkowskian ) , this allows us to compute exactly the leading asymptotic behavior of @xmath1-symbols . * the method of @xmath0-_functions_. we mentioned already that the generating series @xmath59 of is a @xmath0-function in a non - constructive way , but more is actually true . in theorem [ thm.diagonal ] we show that @xmath59 is the diagonal of an explicit rational function . thus , the asymptotics of the taylor coefficients @xmath54 of @xmath59 can be computed in principle via hodge theory , along the lines of @xcite ; see also @xcite . unfortunately , pemantle - wilson s results require genericity assumptions that do not apply even in the case of the @xmath66 or the cube spin network . * the method of wilf - zeilberger . the fundamental theorem of wilf - zeilberger implies that for an arbitrary spin network @xmath2 , the sequence @xmath54 is a multi - dimensional sum of a hypergeometric term , and thus , holonomic , i.e. , it satisfies a recursion relation with coefficients polynomials in @xmath14 , @xcite . the wz method is constructive , and has been implemented in several platforms , @xcite . wz has been implemented for multidimensional sums , but in practice it works best for one or two dimensional sums . using the wz method , we compute the asymptotics of three simple examples , a euclidean @xmath1-symbol , a plane one and a minkowskian @xmath1-symbol , in section [ sec.wz ] . * the ansatz for the asymptotics of hypergeometric multi - dimensional sums . in @xcite the first author developed a constructive ansatz for locating the singularities of the generating series of balanced hypergeometric multi - dimensional sums . the conjecture has been confirmed for one - dimensional sums using the theory of borel transform . for higher dimensional sums the conjecture requires a genericity assumption ( finiteness of the solution set of the so - called variational equations ) which does not seem to hold for spin networks with more than six edges . a proof of the ansatz in the higher dimensional case is expected to follow by developing further the theory of borel transform . the practical limitation of these methods is discussed in section [ sub.challenge ] . in the special case of the tetrahedron graph we present a complete solution to parts ( b ) and ( c ) of problem [ prob.key ] using the method of borel transform . we can explicitly compute the asymptotic expansions for all possible colorings , even in the degenerate and non - physical cases . moreover the quantities in the asymptotic expansion can be expressed as geometric properties of the dual tetrahedron graph interpreted as a metric polyhedron . @xmath67 in this section @xmath68 denotes the tetrahedral spin network colored by an admissible coloring @xmath4 . consider the planar dual graph @xmath69 ( also a tetrahedron ) with the same labelling on the edges . regarding the labels of @xmath69 as edge lengths one may ask whether @xmath70 can be realized as a metric euclidean tetrahedron with these edge lengths . the admissibility of @xmath68 implies that all faces of the dual tetrahedron @xmath69 satisfy the triangle inequality . a well - known theorem of metric geometry @xcite implies that @xmath70 can be realized in _ exactly one _ of three flat geometries * euclidean 3-dimensional space @xmath71 * minkowskian space @xmath72 * plane euclidean @xmath73 . moreover , the above trichotomy is decided by the sign of the _ cayley - menger determinant _ @xmath74 which is a degree six polynomial in the six edge labels ; see definition [ def.cm ] . without loss of generality , we will assume that @xmath68 is _ non - degenerate _ in the sense that all faces of @xmath69 are two dimensional ( i.e. , the triangle inequalities are strict ) . in the degenerate case , the evaluation of @xmath75-symbol is a ratio of factorials , whose asymptotics are easily obtained from stirling s formula . notice that @xmath68 can be degenerate only in the plane and minkowskian cases . every sequence of nilsson type comes with its own numerical invariants : its growth rates , exponents , nilpotency exponents , stokes constants and number field ; see definition [ def.nilsson ] . theorem [ thm.existence ] implies that @xmath1-symbols have such invariants . in this section we will define a set of numerical invariants of tetrahedra and in the next section we will identify those with the numerical invariants of the corresponding @xmath1-symbols . in this section , we will fix a nondegenerate tetrahedral spin network @xmath68 . \(a ) let @xmath76 be the _ exterior dihedral angle _ corresponding to edge @xmath77 of the metric realization of @xmath70 . ( b ) consider the growth rates @xmath78 defined by : @xmath79 ( c ) consider the number field @xmath80 the next lemma , proven in section [ sub.6jdetails ] , explains the top part of figure [ fig.trichotomy ] . the growth rates @xmath81 have the following properties . * @xmath82 * in the euclidean case we have additionally that @xmath83 and @xmath84 and consequently @xmath85 , and @xmath86 . * in the plane case , we have additionally that @xmath87 and consequently @xmath88 . * in the minkowsian case , we have additionally that @xmath89 . exactly one of @xmath78 is inside and one is outside the unit circle . let @xmath90 denote the one inside the unit circle . we define the exponents @xmath91 by @xmath92 * if @xmath4 is euclidean , we define @xmath93 * if @xmath4 is minkowskian , we define @xmath94 * if @xmath4 is plane , we define @xmath95 the definition of @xmath96 and @xmath97 ( homogeneous polynomials in @xmath4 of degree @xmath65 and @xmath98 respectively ) in the plane case can be found in section [ sub.6jdetails ] . in this section we will express the asymptotics of the @xmath1 symbols ( using the minor change @xmath99 in normalization explained in definition [ def.evalu ] ) in terms of the numerical invariants of the previous section . fix a nondegenerate tetrahedron @xmath68 . * if @xmath4 is euclidean or plane , there exist @xmath100 $ ] so that @xmath101 * if @xmath4 is minkowskian , let @xmath102 be the stokes constant corresponding to @xmath90 . there exists @xmath103 $ ] so that @xmath104 moreover , @xmath39 is the associated number field . equation implies in the euclidean case that @xmath105 which is the well - known _ ponzano - regge formula _ for the asymptotics of euclidean @xmath1-symbols . in the euclidean and minkowskian cases these asymptotic formulas were conjectured by ponzano - regge @xcite . an elementary but non - rigorous proof of the euclidean case was given in @xcite . the euclidean case was rigorously proven by @xcite but there was no proof for the minkowksian case . the formula for the plane case is new . notice that exponents @xmath91 in the plane case of theorem [ thm.6j ] is not a half - integer but rather _ the third of an integer_. this phenomenon ( accompanied by the collision of the two points in the plane case of figure [ fig.trichotomy ] ) is natural from the point of view of local monodromy of @xmath0-functions , and is often overlooked in the physics literature . an illustration of theorem [ thm.6j ] in all three geometries is given in section [ sec.wz ] , using the wz method . using theorem [ thm.6j ] , we can compute explicitly the leading asymptotics of an infinite family of spin networks whose underlying graphs can be created from the tetrahedron graph by repeatedly replacing a vertex by a triangle . as follows from the recoupling theory discussed in section [ sec.eval ] the corresponding spin network evaluation will change by multiplication by a certain @xmath1-symbol and a theta . we thus find the asymptotics for all these networks . in particular , this leads to spin networks with an arbitrary number of growth rates , and with associated number fields which are the composite of an arbitrary number of quadratic number fields . in this section we discuss the generating series @xmath106 of a spin network evaluation for a fixed underlying graph @xmath3 . as it turns out , @xmath106 is a rational function . among other things , this implies that the @xmath0-functions of theorem [ thm.g ] come from geometry and more precisely from a variation of mixed hodge structures , much in the spirit of bloch - kreimer @xcite . the next definition introduces the generating series @xmath106 of a fixed graph @xmath3 with cyclic ordering . @xmath106 is a formal power series in commuting variables @xmath107 where @xmath108 , the set of edges of @xmath3 . the _ spin generating function _ @xmath106 of a cubic ribbon graph @xmath3 is defined to be @xmath109\ ] ] where the sum ranges over all colorings @xmath110 of the graph @xmath3 . we will express the spin generating function of an arbitrary graph in terms of the curve polynomial that we will define now . a _ curve _ in a graph @xmath3 is a @xmath65-regular subgraph . by definition the empty set is also a curve . the set of all curves in @xmath3 is called @xmath111 . given a cubic ribbon graph @xmath3 and @xmath112 we define * a polynomial @xmath113 where @xmath114 ( resp . @xmath115 ) when @xmath116 ( resp . @xmath117 ) . * we will call @xmath118 the _ curve polynomial _ of @xmath3 . with the above notation , we define @xmath119 to be the minimal number of intersections when @xmath120 is drawn on the thickened ribbon graph @xmath3 and @xmath121 the next theorem proves the rationality of @xmath106 . for every cubic ribbon graph @xmath3 we have @xmath122 \cap \bq(z_{\e } , e \in e(\ga)).\ ] ] for every spin network @xmath2 , the evaluation @xmath123 is an integer number . @xcite when @xmath3 is planar with the counterclockwise orientation , then @xmath124 unless @xmath120 is empty , in which case @xmath125 . it follows that @xmath126 recovering an earlier theorem by westbury @xcite . the next theorem shows that the generating series @xmath59 of is a _ diagonal _ of a rational function . let us recall the notion of a diagonal . given a power series @xmath127 $ ] and an exponent @xmath128 , we define the @xmath129-diagonal of @xmath130 by @xmath131(f ) z^n \in \bq[[z]]\ ] ] where @xmath132(f)$ ] denotes the coefficient of @xmath133 in @xmath130 . for every spin network @xmath2 we have @xmath134 consequently , the @xmath0-function @xmath59 is the diagonal of a rational function , and thus it comes from geometry ; see lemma [ lem.diagonal ] below . we end this section with the following curious proposition . if @xmath3 is a 2-connected cubic ribbon graph , then the evaluations @xmath135 for all @xmath4 uniquely determine @xmath3 as an abstract trivalent graph . moreover , we only need to use the colors @xmath136 and @xmath115 . in section [ sec.eval ] we discuss the recoupling method for evaluating spin networks which reduces the evaluation of an arbitrary spin network to the @xmath7 and @xmath1-symbols . using this method , and an explicit formula for the @xmath7 and @xmath1-symbols , we evaluate the drum and the @xmath66 graphs . in section [ sec.hyper ] we prove that the evaluation of a spin network is given by a multi - dimensional sum of a balanced hypergeometric term . this , together with the results of @xcite , implies theorem [ thm.g ] and theorem [ thm.existence ] . in section [ sec.borel ] we apply techniques of borel transform ( developed in @xcite ) that give exact formulas for the leading asymptotic behavior of one - dimensional sums of hypergeometric terms . since @xmath1-symbols are given by such sums , this allows us to compute exactly the leading asymptotic behavior of @xmath1-symbols . section [ sec.6j ] identifies the combinatorial leading asymptotic behavior with geometric quantities associated to the metric tetrahedron in all cases , euclidean , plane or minkowskian . in section [ sec.wz ] we present an alternative constructive method to compute the asymptotic behavior of evaluations of spin networks , namely the wz method . in section [ sec.chromatic ] we discuss in detail the chromatic evaluation of spin networks and obtain the integrality and rationality results of the spin generating series . the results were conceived in a workshop in aarhus , denmark , and presented in hanoi , vietnam and strasbourg , france in the summer of 2007 . the authors wish to thank the organizers for their hospitality . s.g . wishes to thank d. zeilberger for many enlightening conversations . we start by recording some elementary facts about spin network evaluations . let @xmath137 be a spin network . 1 . inserting a vertex colored @xmath138 in the interior of an edge of @xmath3 colored @xmath77 does not change the evaluation of the spin network . 2 . changing the cyclic ordering at a vertex whose edges are colored @xmath77 , @xmath139 , @xmath140 changes the evaluation by a sign @xmath141 \(a ) the chosen normalization introduces an extra factor @xmath142 for the new vertex labeled @xmath138 , while it follows from the definition that one also inserts an extraneous summation over permutations in the pre - existing edge labeled @xmath77 . since @xmath143 the evaluation is unchanged . ( b ) changing the cyclic order at a vertex with edge labels @xmath10 has the following effect . the alternating sum at each of the adjacent edges is multiplied by the permutation that turns the arcs in the edge by 180 degrees . this element has sign @xmath144 in @xmath145 . as a consequence of part ( a ) of the above lemma , an edge labeled @xmath136 in a spin network can be removed without affecting the evaluation . there is an alternative _ bracket normalization _ @xmath146 of the evaluation of a spin network @xmath2 which agrees with a specialization of the _ jones polynomial _ and the _ kauffman bracket_. the _ bracket normalization _ of a spin network @xmath2 is defined by @xmath147 where @xmath148 this normalization has the property that it coincides with the kauffman bracket ( jones polynomial ) of a quantum spin network evaluated at @xmath149 @xcite . however , @xmath146 is not necessarily an integer number , and the analogous generating series does not satisfy the rationality property of theorem [ thm.gf ] . ( of proposition [ prop.uniqued ] ) for convenience we work with the unnormalized evaluation @xmath12 . a 2-connected graph is determined up to isomorphism by its 2-regular subgraphs , simply take a polygon for every 2-regular subgraph and identify equal edges @xcite . to decide whether a set @xmath120 of edges forms a 2-regular subgraph , let @xmath4 be the coloring that assigns @xmath115 to members of @xmath120 and @xmath136 to the other edges . by definition of the spin network evaluation @xmath120 is an @xmath14 component 2-regular subgraph if and only if @xmath150 . in this subsection we describe a way of evaluating spin networks by recoupling . we will reduce the evaluation of spin networks to multi - dimensional sums of @xmath1 and @xmath7-symbols . the value of the @xmath1 and @xmath7-symbols is given by the following lemma of @xcite and @xcite , using our normalization . \(a ) let @xmath68 denote a tetrahedron labeled and oriented as in figure [ fig.exspin ] or figure [ fig.tandd ] ( left ) with @xmath151 . its normalized evaluation is given by @xmath152 where , as usual @xmath153 denotes the multinomial coefficient when @xmath154 , and @xmath155 are the half sums of the labels in the three quadrangular curves in the tetrahedron and @xmath156 are the half sums of the thee edges emanating from a given vertex . in other words , the @xmath155 and @xmath156 are given by @xmath157 @xmath158 ( b ) let @xmath159 denote the @xmath9 spin network of figure [ fig.exspin ] admissibly colored by @xmath160 . then we have @xmath161 note that the value of a @xmath14-colored unknot is by lemmas [ lem.evaluation ] and [ lem.6j3jvalue ] equal to @xmath162 . recoupling is a way to modify a spin network locally , while preserving its evaluation . this is done as in the topmost picture in figure [ fig.recoupling ] . this formula is called the recoupling formula and follows from the recoupling formula in @xcite , using our conventions . the other two pictures in the figure show two important special cases of the recoupling formula , the bubble formula and the triangle formula . @xmath163 the bubble formula shown on the left of figure [ fig.recoupling ] serves to remove all bigon faces . likewise the triangle formula can be used to remove triangles . to see why the recoupling formula suffices to write any spin network as a multi sum of products of @xmath1-symbols divided by thetas we argue as follows . applying the recoupling formula to a curve in the graph reduces its length by one . keep going until you get a multiple edge which can then be removed by the bubble formula . although the triangle formula follows quickly from the bubble formula and the recoupling formula it is important enough to state on its own . this formula shows that the evaluation of the class of _ triangular networks _ is especially simple . the triangular networks , mentioned in remark [ rem.6j ] are the planar graphs that can be obtained from the tetrahedron by repeatedly replacing a vertex by a triangle . by the triangle formula any triangular network is simply a product of @xmath1-symbols divided by thetas . to illustrate how recoupling theory works , let us evaluate the regular @xmath164-sided drum and @xmath66 . consider the @xmath164-sided drum network as shown in equation ( for @xmath165 ) where every edge is colored by the integer @xmath14 . in the figure we have left out most of the labels @xmath14 for clarity . by convention unlabeled edges are colored by @xmath14 . performing the recoupling move on every inward pointing edge we transform the drum into a string of bubbles that is readily evaluated . @xmath166 observing that if @xmath14 is odd the network is not admissible ( and thus evaluates to zero ) , and denoting the tetrahedron and the theta with one edge colored by @xmath167 and the others by @xmath14 by @xmath168 and @xmath169 we conclude the following formula for the @xmath14-colored @xmath164-sided drum . if @xmath170 is even we have @xmath171 and if @xmath14 is odd we have @xmath172 . for small values of @xmath164 the drum can be evaluated in a more straight forward way , thus providing some well known identities of @xmath1-symbols . namely when @xmath173 we get zero , when @xmath174 we find some thetas and when @xmath175 we have by the triangle formula a product of two @xmath1-symbols thus giving a special case of the biedenharn - elliott identity @xcite . for @xmath176 we find a formula for the regular cube , we know of no easier evaluation in this case . a similar computation for @xmath66 cyclically ordered as a plane hexagon with its three diagonals gives the following . if @xmath170 is even we have @xmath177 and if @xmath14 is odd we have @xmath178 . note the similarity between @xmath179 and @xmath66 . the only difference is the sign that comes up in the calculation when one needs to change the cyclic order . the extra sign makes @xmath180 for all odd @xmath181 . this is because changing the cyclic ordering at a vertex takes the graph into itself , while it produces a sign @xmath182 when all edges are colored @xmath183 . in this section we prove the existence of an asymptotic expansion for all spin networks using the theory of @xmath0-functions and the combinatorial fact that any spin network can be evaluated as a multi - dimensional sum of a balanced hypergeometric term . an @xmath184-dimensional _ balanced hypergeometric term _ @xmath185 ( in short , _ balanced term _ ) in variables @xmath186 , where @xmath49 and @xmath187 , is an expression of the form @xmath188 where @xmath189 are algebraic numbers for @xmath190 , @xmath191 for @xmath192 , and @xmath193 are integral affine linear forms in @xmath186 that satisfy the _ balance condition _ @xmath194 we assume that the convex polytope @xmath195 in @xmath196 defined by @xmath197 for @xmath192 and @xmath198 is compact . for every spin network @xmath2 there exists a balanced term @xmath199 with @xmath200 such that @xmath201 for every @xmath49 . using the recoupling formulae from the previous paragraph we can write @xmath45 as a multi - dimensional sum of products of @xmath1-symbols , @xmath7-symbols and unknots ( i.e. , @xmath202-symbols ) with a denominator consisting of @xmath7-symbols . it follows from equations and that the @xmath1-symbols ( resp . @xmath7-symbols ) are balanced 1-dimensional ( resp . @xmath136-dimensional ) sums , thus the ratio of the product of the @xmath7-symbols by the product of the @xmath7-symbols is a balanced multi - dimensional sum . the unknots can be written as @xmath203 and are therefore balanced as well . it is easy to check that admissibility guarantees that the multi - dimensional sum has finite range . beware that the term @xmath199 constructed in the proof of lemma [ lem.fgc ] is neither unique nor canonical in any sense . lemma [ lem.fgc ] and the following result implies theorem [ thm.existence ] . * thm.2 ) ( a ) if @xmath29 is a multi - dimensional sum of a balanced term , then the generating series @xmath204 is a @xmath0-function . ( b ) if @xmath205 is a @xmath0-function , its taylor series at @xmath56 is a sequence of nilsson type . this argument also implies that the sequences @xmath206 and @xmath20 are also asymptotic to a rational power of @xmath207 times a sequence of nilsson type . this follows from the fact that the above normalizations of the spin networks differ from definition [ def.eval ] by a ratio of factorials of linear forms in @xmath14 , which can be asymptotically expanded using stirling s formula . let us end this section with a lemma which shows that diagonals of rational functions are @xmath0-functions that come from geometry . this is relevant to spin networks since we know from theorem [ thm.diagonal ] that the generating function @xmath208 for the evaluations of a spin network is a diagonal of a rational function . fix a power series @xmath127 $ ] convergent at the origin and an exponent @xmath128 and consider the diagonal @xmath209 $ ] as in definition [ def.diagonal ] . let @xmath210 denote a small @xmath184-dimensional torus around the origin . then we have the following . with the above assumptions , we have @xmath211 with the notation of definition [ eq.diagonal ] , an application of cauchy s theorem gives for every natural number @xmath14 @xmath212(f)=\frac{1}{(2 \pi i)^r } \int_{\calc } \frac{f(x_1,\dots , x_r)}{(x_1 \dots x_r)^{n+1 } } dx_1 \wedge \dots \wedge dx_r.\ ] ] summing up for @xmath14 and interchanging summation and integration concludes the proof . if in addition @xmath213 is a rational function , then the singularities of the analytic continuation of the right hand - side of can be analyzed by deforming the integration cycle @xmath210 and studying the corresponding variation of mixed hodge structure as in @xcite . such @xmath0-functions come from geometry ; see @xcite . the results of section [ sec.hyper ] are powerful but non - constructive . in this section we will use constructive results from the theory of borel transform , known in the case of balanced terms of rank @xmath115 . these results appear in @xcite , and in the present paper we will quote them without proofs . let us specialize to 1-dimensional balanced terms @xmath214 , that is expressions of the form @xmath215 satisfying the balancing condition . the relevant example of these is the @xmath1-symbol , equation . fixing a 1-dimensional balanced term @xmath214 , summation over @xmath167 ( it is a finite sum ) gives rise to a sequence @xmath216 to describe and compute the asymptotics of such sequences we need the following notation . for affine linear forms we use the notation @xmath217 and @xmath218 . next we define the multivalued analytic functions @xmath219 @xmath220 and @xmath221 are analytic functions on the open interval @xmath222 ( which is determined by @xmath223 ) and they have single - valued analytic continuation in the _ doubly cut _ complex plane @xmath224 \cup [ m , \infty))$ ] . @xmath220 and @xmath221 can be further analytically continued as multivalued functions in the complex plane minus a finite set of points @xmath225 . define the constants @xmath226 the set of critical points of @xmath220 satisfies the variational equation @xmath227 if @xmath228 is a critical point of @xmath220 , then @xmath229 where @xmath230 let @xmath39 denote the number field generated by the solutions of the variational equation . suppose that @xmath228 is a critical point of order @xmath231 for @xmath220 , that is @xmath232 for @xmath233 and @xmath234 and @xmath221 is non - vanishing at @xmath228 , i.e. , @xmath235 . let @xmath236 and let @xmath237 and @xmath238 we will require a modification of the above definition of the stokes constant when @xmath228 lies in one of the cuts @xmath239 \cup [ m,\infty)$ ] . in that case , we define @xmath240 with the above assumptions we have @xmath241 where @xmath242 $ ] is gevrey-1 and @xmath228 runs through a subset of solutions of the variational equation . the selected subset can be described algorithmically and in case of the @xmath1-symbols it is discussed below . the purpose of this section is to prove and explain theorem [ thm.6j ] on the explicit expansions of the @xmath1-symbols . since the asymptotics are phrased in terms of geometric properties of the tetrahedron dual to the spin network , we start with a section on the geometry of the tetrahedron . next we apply theorem [ thm.res ] to find the asymptotic expansion and finally we work out the details in each of the geometric cases . the asymptotics of the @xmath1-symbols turn out to be intimately related to the geometry of the planar dual tetrahedron . from now on , we will assume that @xmath4 is non - degenerate , i.e. that the faces of @xmath70 are two dimensional triangles . recall that the @xmath1-symbol is a tetrahedral spin network @xmath68 admissibly labeled as in figure [ fig.exspin ] with @xmath151 . its dual tetrahedron @xmath70 is also labeled by @xmath4 . the tetrahedron and its planar dual , together with an ordering of the vertices and a coloring of the edges of the dual is depicted in figure [ fig.tandd ] . when a more systematic notation for the edge labels is needed we will denote them by @xmath243 ( or @xmath244 if there can not arise any confusion ) it follows that @xmath245 @xmath246 recall from section [ sub.g6j ] that we interpret the labels of the dual tetrahedron @xmath69 as edge lengths in a suitable flat geometry . one may now ask for a condition that allows one to realize @xmath70 in a flat metric space such that the edge lengths equal the edge labels . labeling the vertices of @xmath70 as in figure [ fig.tandd ] we can formulate such a condition in terms of the cayley - menger determinant . this is a homogeneous polynomial of degree @xmath98 in the six variables @xmath247 . we give a definition of the determinant following @xcite . given numbers @xmath243 define the ( modified ) _ cayley - menger _ matrix @xmath248 by @xmath249 for @xmath250 and @xmath251 for when @xmath252 or @xmath253 . in terms of the coloring @xmath151 of a tetrahedron , we have @xmath254 we can now state the criterion for the realizability of @xmath70 . for a proof see @xcite or @xcite . consider a non - degenerate tetrahedron @xmath70 . the sign of the cayley - menger determinant determines in what space the tetrahedron can be realized such that the edge labels equal the edge lengths * if @xmath255 then the tetrahedron is realized in euclidean space @xmath71 . * if @xmath256 then the tetrahedron is realized in the euclidean plane @xmath73 . * if @xmath257 then the tetrahedron is realized in minkowski space @xmath72 . in each case the volume of the tetrahedron is given by @xmath258 . the 6 dimensional space of non - degenerate tetrahedra thus consists of regions of minkowskian and regions of euclidean tetrahedra . it turns out to be a cone that is made up from one connected component of three dimensional euclidean tetrahedra and two connected components of minkowskian tetrahedra . the three dimensional euclidean and minkowskian tetrahedra are separated by plane tetrahedra . the plane tetrahedra also form two connected components , representatives of which are depicted in figure [ fig.trichotomy ] . the tetrahedra in the plane component that look like a triangle with an interior point are called _ triangular _ and the plane tetrahedra from that other component that look like a quadrangle together with its diagonals are called _ quadrangular_. we use the same names for the corresponding minkowskian components . an integer representative of the triangular plane tetrahedra is not easy to find as the smallest example is @xmath259 . we finish this section with a short discussion of the dihedral angles of a tetrahedron realized in either of the three above spaces . the cosine and sine of these angles can be expressed in terms of certain minors of the cayley - menger matrix . define the _ adjugate matrix _ @xmath260 whose @xmath244 entry is @xmath261 times the determinant of the matrix obtained from @xmath262 by deleting the @xmath263-th row and the @xmath264-th column . define @xmath265 to be the matrix obtained from @xmath262 by deleting both the @xmath263-th row and column and the @xmath264-th row and column . the law of sines and the law of cosines are well - known formulae for a triangle in the euclidean plane . the next lemma is the _ law of sines _ and the _ law of cosines _ for a tetrahedron in all three flat geometries . let @xmath266 be the exterior dihedral angle at the opposite edge @xmath244 . the following formula is valid for all non - degenerate tetrahedra @xmath267 see @xcite for a proof in the euclidean and plane case . the minkowskian case is entirely analogous once one uses the right conventions on angles @xcite . given a minkowskian tetrahedron the angle between two outward pointing normal vectors has the form @xmath268 , where @xmath269 and @xmath270 . since the normal vectors are time - like they can either be in the same sheet or in different sheets of the unit hyperboloid . if they are in the same sheet then @xmath271 is defined to be the length the hyperbolic line between them and @xmath272 . if they are in opposite sheets then @xmath273 is the length of the line between one and the reflection of the other in the horizontal plane and @xmath274 . we now start to apply theorem [ thm.res ] to determine the asymptotics of the non - degenerate @xmath1-symbol . in this section it is convenient to use the following normalization @xmath275 for spin networks . we define the _ unitary normalization _ @xmath275 of a spin network evaluation @xmath2 to be @xmath276 where @xmath277 and @xmath278 are the colors of the edges at vertex @xmath25 . as it turns out , with this normalization the asymptotics of the @xmath1-symbols matches the corresponding geometry well . since the theta is a quotient of factorials one can use stirling s formula @xcite to find its asymptotics and hence one can see how the new choice in normalization effects the asymptotic expansion . for later use we record the asymptotic expansion of the normalization factor in the case of a @xmath1-symbol . let us define @xmath279 @xmath280 let @xmath4 be an admissible coloring of a tetrahedron . then we have @xmath281 where @xmath282 and @xmath283 for all @xmath284 . thus , the asymptotics of @xmath285 , is reduced to the study the asymptotics of @xmath286 . since this is a 1-dimensional sum of a balanced term given by equation , we can apply theorem [ thm.res ] to find its asymptotic expansion . equation implies that the balanced term is given by @xmath287 with the notation of theorem [ thm.res ] we have @xmath288 , and the linear forms @xmath289 are given by @xmath290 @xmath291 and @xmath292 , @xmath293 . it follows that @xmath294 the functions @xmath220 and @xmath221 from theorem [ thm.res ] in this case are @xmath295 @xmath296 @xmath220 and @xmath221 are analytic functions on the open interval @xmath297 and they have single - valued analytic continuation in the doubly cut complex plane @xmath298 \cup [ \min\ { s_i\ } , \infty))$ ] . @xmath220 and @xmath221 can be further analytically continued as multivalued functions in @xmath299 . the possible critical values of @xmath300 that determine the growth rate are given by @xmath301 here @xmath25 is a solution of the variational equation which is the equation @xmath302 at first sight , the above equation appears to be a quartic equation for @xmath228 . however , the coefficients of @xmath303 and @xmath304 is zero because @xmath305 . thus , equation is quadratic in @xmath228 , with coefficients which are polynomial in @xmath306 . it is convenient to introduce a fourth number @xmath307 . we can then write the variational equation in the form @xmath308 where @xmath96 , @xmath309 and @xmath310 are homogeneous polynomials in @xmath4 of degrees @xmath65 , @xmath98 and @xmath311 respectively . a calculation gives that @xmath312 the next lemma identifies the number field generated by the variational equation in terms of the cayley - menger determinant . let @xmath313 be the discriminant of the variational equation . then we have @xmath314 where @xmath262 is the cayley - menger matrix . the number field @xmath39 generated by the solutions to the variational equation is given by @xmath315 the proof follows from an explicit calculation , keeping in mind that @xmath313 and @xmath74 are homogeneous polynomials in @xmath4 of degree @xmath75 . the next lemma determines the order of vanishing of @xmath220 at the solutions of equation . let @xmath313 be the discriminant of the variational equation . * the tetrahedron is degenerate iff there is a solution to the variational equation that equals @xmath155 , or @xmath156 for some @xmath316 . * assume the tetrahedron is non - degenerate . the two solutions @xmath317 to the variational equation satisfy @xmath318 * if in addition @xmath319 then the two solutions coincide , @xmath320 and we have @xmath321 where @xmath96 is given in . * in the euclidean or plane case , @xmath322 does not lie in the double cut @xmath239\cup[m,\infty)$ ] . in the minkowskian case , @xmath322 are real and exactly one of them ( call it @xmath25 ) lies in the double cut @xmath239\cup[m,\infty)$ ] . in the proof it is convenient to define @xmath307 so that all indices run from @xmath115 to @xmath311 . this convention will only be used in this proof . ( a ) if the tetrahedron is degenerate then there are @xmath316 such that @xmath323 . plugging in this value for @xmath228 will set both terms in the equation to zero as required . on the other hand if @xmath324 is a solution to the variational equation then @xmath325 , so that @xmath326 for some @xmath263 and likewise when @xmath327 . \(b ) note that @xmath328 since all edge labels are positive so that a solution always exists . by ( a ) we know that the solutions to the variational equation are not equal to @xmath155 or @xmath156 . by differentiating @xmath220 we find that @xmath329 . using the variational equation @xmath330 we can write @xmath331 we see that @xmath332 . now choose the solutions such that @xmath333 so that @xmath334 . \(c ) since @xmath335 the solutions @xmath322 coincide and will be written as @xmath25 . differentiating we find @xmath336 . on the other hand @xmath319 and we assumed that our tetrahedron is non - degenerate , so by part ( a ) @xmath337 . therefore @xmath338 adding these to similar equations for @xmath339 implies @xmath340 multiplying through by the variational equation @xmath341 we get @xmath342 where @xmath96 is given in equation . \(d ) the euclidean or plane case follows from ( b ) . for the minkowskian case , there are connected components ( types ) of minkowskian tetrahedra : the rectangular type where @xmath343 and the _ triangular type _ where we have @xmath344 where @xmath345 is the distinguished triangle of the triangular tetrahedron . this fact is in turn proven by continuity . it is true by explicit computation ( using the wz method of section [ sec.wz ] ) for the two examples @xmath346 of rectangular type and triangular type respectively . the result follows by continuity for all minkowskian nondegenerate tetrahedra , since @xmath347 only for degenerate tetrahedra . from now on we will always assume that our @xmath1-symbols are non - degenerate . using lemma [ lem.lambda ] and theorem [ thm.res ] , it follows that the possible growth rates in the asymptotic expansion of @xmath348 are given by @xmath349 we close this subsection with a lemma connecting these growth rates to the dihedral angles in the tetrahedron . let @xmath76 be the dihedral angle corresponding to the edge labeled @xmath77 as introduced in section [ sub.tetrahedron ] . 1 . the growth rate @xmath78 satisfies @xmath350 2 . moreover , @xmath351 \(a ) the strategy is to collect equal powers in @xmath78 . it follows that we can write @xmath78 as a product @xmath352 where we define @xmath353 and @xmath354 now we will show that @xmath355 . since @xmath356 and @xmath357 we can clear denominators in the expression for @xmath358 from lemma [ lem.angles ] and the expression for @xmath359 . equality of the numerators is then shown by direct calculation using equation . equation follows . \(b ) to show equation , first note that @xmath360 and hence @xmath361 in the product of the @xmath362 we can also collect like terms together in the form @xmath363 in the denominator . we get @xmath364 using the variational equation and @xmath355 proves part ( b ) . lemmas [ lem.lambdatheta ] and [ lem.vareq ] prove lemma [ lem.lambdas ] regarding the properties of the growth rates . combining our results , we can finally prove theorem [ thm.6j ] by reading off the numerical invariants from theorem [ thm.res ] . first of all the growth rates @xmath78 are identified with the dihedral angles by lemma [ lem.lambdatheta ] . in the euclidean and the plane case both rates contribute , whereas in the minkowskian case we include only the contribution of @xmath90 , corresponding to the distinguished solution @xmath25 of the variational equation that lies in the cut ; see part ( d ) of lemma [ lem.vareq ] . in the terminology of section [ sub.tetrahedron ] and lemma [ lem.vareq ] we have @xmath365 in the triangular component and @xmath366 in the quadrangular component , and always @xmath367 . next the exponent in the leading power of @xmath14 is given by equation and the order of vanishing is determined by lemma [ lem.vareq ] . we have @xmath368 in the euclidean and minkowskian cases and @xmath369 in the plane cases . lemma [ lem.field ] identifies the number field generated by the solutions to the variational equation with @xmath39 from definition [ def.6jgrates ] . finally we consider the stokes constants in all three cases . the stokes constant is computed from theorem [ thm.res ] using lemmas [ lem.lambdatheta ] and [ lem.vareq ] . in the euclidean case we have to refine lemma [ lem.lambdatheta ] ( b ) to choose the right square root of the left hand side . this is done by considering the simple example of the regular tetrahedron and extending by continuity . we thus find @xmath370 now applying part ( b ) of lemma [ lem.vareq ] and lemma [ lem.field ] , we can rewrite the above as @xmath371 in the minkowskian case we have to choose the stokes constant corresponding to @xmath90 . this is computed as in the euclidean case except that the solutions to the variational equation lie in the cut . we therefore use equation so that we take absolute values everywhere . this explains the sign in . it makes the stokes constant positive by canceling the real part of the minkowskian angle sum , which is @xmath372 in the triangular case and @xmath373 in the quadrangular case . finally in the plane case we have by lemma [ lem.vareq ] that @xmath374 and we know that @xmath337 . furthermore @xmath375 , so @xmath376 where we used part ( c ) of lemma [ lem.vareq ] and the fact that the dihedral angles are real . since the tetrahedron is plane , we have @xmath377 , so @xmath378 with the notation of . therefore we can take out the @xmath379 terms to get @xmath380 recall from equations , and that @xmath156 , @xmath96 and @xmath309 are homogeneous polynomials of @xmath4 . a calculation shows that @xmath381 is a homogeneous polynomial of degree @xmath98 . it is unclear to us what is the geometric interpretation of the rational function @xmath382 . in this section we will compute the asymptotic expansion of three representative @xmath1-symbols using the wz method . the main observation is that a @xmath1-symbol is given by one - dimensional sum of a balanced hypergeometric term ; see equation and definition [ def.hyperg ] . the fundamental theorem of wilf - zeilberger implies that such sums are holonomic , i.e. , they satisfy linear recursion relations with coefficients polynomials in @xmath14 . the wz method is constructive and works well for one - dimensional sums , implemented in several platforms , see @xcite . given a recursion relation for the evaluation of the @xmath1-symbols , together with the fact that it is a sequence of nilsson type , allows us to compute explicitly its asymptotic expansion . the three examples we consider are the simplest representatives of the three geometric types of @xmath1-symbols . let @xmath383 and consider the sequences @xmath384 and @xmath385 defined by @xmath386 the command @xmath387 loads the package of @xcite into mathematica . the command @xmath388 defines the summand of the sequence @xmath29 , and the command @xmath389 computes the following second order linear recursion relation for the sequence @xmath29 @xmath390 this linear recursion has two formal power series solutions of the form @xmath391 where @xmath392 are two complex numbers of absolute value @xmath115 . these growth rates @xmath78 computed with the wz method match those of theorem [ thm.6j ] . indeed , the theorem states that @xmath393 an algebraic calculation confirms that @xmath394 the coefficients of the formal power series @xmath395 are in the number field @xmath396 and the cayley - menger determinant is @xmath397 confirming lemma [ lem.field ] and theorem [ thm.6j ] . more is actually true . namely , the sequence @xmath29 generates two new sequences @xmath398 and @xmath399 defined by @xmath400 where @xmath401 . each of the sequences @xmath402 are factorially divergent . however , the generating series @xmath403 are @xmath0-functions ( as follows from @xcite ) , and the sequences @xmath404 are of nilsson type , with exponential growth rates @xmath405 . the asymptotics of each sequence @xmath404 gives rise to finitely many new sequences , and so on . all those sequences span a finite dimensional vector space , canonically attached to the sequence @xmath29 . this is an instance of resurgence , and is explained in detail in ( * ? ? ? * sec.4 ) . the second order recursion relation for the plane and the minkowskian examples has lengthy coefficients , and leads to the following sequences @xmath406 and @xmath407 @xmath408 where in the plane case we have @xmath409 and in the minkowskian case we have @xmath410 again in agreement with theorem [ thm.6j ] . the purpose of this section is to prove the rationality statement of the spin generating function stated in theorem [ thm.gf ] of section [ sec.gf ] . we will use the _ chromatic evaluation method _ which goes back to @xcite . our proof builds on earlier work by @xcite and @xcite on planar spin networks . the chromatic method for the evaluation of a spin network is based on introducing a generalized evaluation @xmath411 following the algorithm of definition [ def.evalp ] , with the exception that a collection of @xmath14 loops is assigned the value @xmath412 instead of @xmath13 . therefore , @xmath413 now the crucial observation is that @xmath414 is a polynomial in @xmath181 with integer coefficients , so that @xmath415 is determined once we know the values of @xmath416 for all positive @xmath181 . for positive @xmath181 there exists an alternative interpretation of @xmath416 as a contraction of tensors . for convenience we will orient the edges of @xmath3 . let @xmath220 be an @xmath181 dimensional vector space and for @xmath417 define the following permutation operators @xmath418 , @xmath419 . at every edge colored @xmath14 we can now speak of the map @xmath420 . note that the single arcs in the expanded diagram for @xmath2 now correspond to copies of @xmath220 . the systems of arcs at the vertices indicate how to contract ( compose ) the tensors on the edges . since all the tensors involved are linear combinations of kronecker deltas we can for each choice of permutations count the number of contracted deltas and this will exactly contribute a term @xmath421 for some @xmath167 since the trace of a delta is simply @xmath181 . this coincides with the counting and evaluation of loops in every state , since a contracted delta represents a loop . however there is also another way to carry out this tensor calculation . we can also calculate @xmath416 by writing out the contraction of tensors into the single contributions of basis elements in the various tensor products . this will yield an explicit formula for the evaluation of a general spin network in terms of so called curve configurations on @xmath3 . recall that a curve is a @xmath65-regular subgraph of @xmath3 and a curve configuration is a function from the set @xmath262 of all curves to the natural numbers , @xmath422 . for such functions we define @xmath423 to be the sum of the function values . define sign @xmath424 , where for any subset @xmath425 is the minimal number of intersections one can have when drawing @xmath426 copies of @xmath140 for every curve on the thickened ribbon graph @xmath3 . for every spin network @xmath2 we have @xmath427 this implies in particular that the generating series @xmath106 of definition [ def.sg ] satisfies @xmath428 $ ] proving one part of theorem [ thm.gf ] . ( of lemma [ lem.binomials ] ) we will calculate @xmath416 by writing out the contraction of tensors into the single contributions of basis elements in the various tensor products . given a choice of permutations we place tensor products of basis elements @xmath429 on every edge . such a term contributes @xmath136 if some basis vectors on the same strand are unequal or the same basis element occurs twice in a tensor product corresponding to an edge . otherwise the contribution is always @xmath430 . to enumerate the non - zero states in an orderly fashion including the signs of their contributions we proceed in three stages . first we choose which sets of arcs get labeled by the same base vector . then we assign base vectors to the loops in the curves . finally we count the number of such labeled states . in the first stage we choose the collections of all loops that are labeled by the same base vector . for non - zero states these collections are necessarily @xmath65-regular subgraphs of @xmath3 , also known as curves . we will view curves both as sets of the loops they contain and as subgraphs of @xmath3 . this first stage of the calculation is summarized in the concept of a curve configuration . to every curve @xmath431 we assign the number of distinct base vectors it occurs with . the curve configuration @xmath432 determines the total coloring of the spin network @xmath4 . however there may be many curve configurations yielding the same coloring @xmath4 . note that all states with the same curve configuration contribute the same sign . this is because of planarity of the ( multiple ) annulus neighborhood of the curve : two loops in the same curve can only cross an even number of times . the common sign is determined by the cyclic ordering of the graph and equals @xmath424 . this follows directly from a count of the number of intersections of such states . in the second stage we choose base vectors and we assign them to the loops such that all the loops with the same base vector constitute exactly one copy of a curve . first we need to choose which base vectors to use . there are @xmath433 ways to do this . then we need distribute these over the various curves , assigning @xmath426 base vectors to the curve @xmath140 in no particular order . there are @xmath434 ways to do this . in the last stage we need to count how many arrangements of loops are possible given a curve configuration with base vectors at every loop . as was shown in @xcite there are @xmath435 ways to do this . their argument does not make use of planarity but for completeness sake we still reproduce the argument here . if we ignore the labelling by base vectors we will count exactly @xmath436 of the possible states . however if we also choose to ignore the fact that in contributing states the loops can not self intersect , then counting possibly zero states will multiply our count by the same factor @xmath437 . to see why , note that to modify a zero state takes one permutation per curve . recall that @xmath435 corresponds to a choice of a permutation on the arcs on each side of every vertex in the ribbon graph . we will show that there are exactly @xmath435 ( possibly zero ) states corresponding to a given curve configuration . first choose one such a state @xmath102 . we will show that all others can be reached by inserting a unique permutation at every side of every vertex , with the appropriate modification . at any rate there can not be more than @xmath435 such states , because arcs belonging to one side of a vertex can never be permuted to become part of another side , because that would destroy the curve configuration . even inserting a permutation at the side of a vertex may destroy the curve configuration , but in this case this can be compensated as follows . choose an edge adjacent to the chosen side of the vertex and insert the inverse of the part of the permutation swapping arcs belonging to different curves . exactly because of the fact that this compensating permutation involves only arcs corresponding to unequal curves this does not interfere with any other inserted permutations at the sides of vertices or their compensations . summarizing , we get @xmath438 and rescaling we get @xmath439 finally setting @xmath440 provides a general formula for spin network evaluation as given in the lemma . to find a generating function for these evaluations , we first introduce a more general generating function . define the _ curve generating function _ @xmath441 of a graph @xmath3 by @xmath442 where @xmath441 is a function of @xmath443 variables , one for each curve in @xmath3 . the exponents of these variables encode a curve configuration , and the evaluation @xmath444 of a curve configuration @xmath445 is given by latexmath:[\[\la\mathcal{l}\ra = ( -1)^{\i(\mathcal{l})+\|\mathcal{l}\| } ( the spin generating function @xmath106 can be recovered from the curve generating function @xmath441 by substituting @xmath447 , because the color of an edge is the sum of the number of curves ( with multiplicity ) containing that edge . ( of theorem [ thm.gf ] ) our goal now is to write @xmath441 and hence @xmath106 as a rational function . as a first step we note that if we drop the term @xmath424 then we can sum the resulting series using the binomial series and the multinomial theorem to get @xmath448 using this together with the fourier expansion of @xmath424 viewed as a function on the group @xmath449 as @xmath450 we find @xmath451 where @xmath452 and @xmath453 . in this section we list some challenges and future directions . recall from definition [ def.radius ] that every spin network @xmath2 ( and consequently every ribbon graph ) has a spectral radius @xmath58 and a number field @xmath60 . our first problem implies the existence of an upper bound of the spectral radius in terms of @xmath3 alone . this is phrased more conveniently in terms of the unitary evaluation of definition [ def.evalu ] . this problem may be proven using unitarity and locality in a way similar to the proof that the reshetikhin - turaev invariants of a closed 3-manifold grow at most polynomially with respect to the level ; see ( * ? ? ? * thm.2.2 ) . recall the spectral radius of a cubic ribbon graph from definition [ def.radius ] ; it is a nonnegative algebraic number . our next problem is a version of the volume conjecture for classical spin networks with all edges colored by @xmath65 . theorem [ thm.6j ] confirms this problem for the following ribbon graphs : the @xmath9 , the tetrahedron , the 3-faced drum and more generally for the infinite family of ribbon graphs of remark [ rem.6j ] . there is convincing numerical evidence for the above problem in the case of the @xmath66 and the cube . at the moment , we do not know the answer to problem [ prob.3 ] for the cube with all edges colored by @xmath65 . numerical evidence suggests that there are @xmath457 equally spaced singularities in the boundary of the disk of convergence of @xmath458 . prove that for every coloring @xmath4 of the tetrahedron spin network @xmath68 , the sequence @xmath462 satisfies a second order recursion relation with coefficients polynomials in @xmath14 . can you compute the coefficients of this recursion from @xmath4 alone ? our last problem concerns an integrality property of quantum spin networks . our paper discusses four normalizations @xmath12 , @xmath135 , @xmath463 and @xmath464 for the evaluation of a classical spin network @xmath2 ; see definitions [ def.evalp ] , [ def.eval ] , [ def.evalb ] and [ def.evalu ] . with the notation of these definitions , these normalizations are related by @xmath465 on the other hand , if @xmath2 is a quantum spin network , the kauffman bracket evaluation @xmath466 makes sense , and in general it is a rational function of the bracket parameter @xmath96 ; see definition [ def.evalb ] and @xcite . the following problem is a first step towards the categorification of the evaluation of a quantum spin network . let us end this section with a remark . the main results of our paper can be extended to evaluations of spin networks corresponding to higher rank lie groups . this will be discussed in a later publication . kirillov and n.yu . reshetikhin , _ representations of the algebra @xmath469-orthogonal polynomials and invariants of links _ , in infinite - dimensional lie algebras and groups , adv . * 7 * ( 1989 ) 285339 . e. suarez - peiro , _ a schlfli differential formula for simplices in semi - riemannian hyperquadrics , gauss - bonnet formulas for simplices in the de sitter sphere and the dual volume of a hyperbolic simplex _ , pacific j. math . * 194 * ( 2000 ) 229255 .	a spin network consists of a cubic ribbon graph ( i.e. , an abstract trivalent graph with a cyclic ordering of the vertices around each edge ) and an admissible coloring of its edges by natural numbers . the evaluation of a spin network , with our normalization , is an integer number . the main results of our paper are : ( a ) an existence theorem for the asymptotics of evaluations of arbitrary spin networks ( using the theory of @xmath0-functions ) , ( b ) a rationality property of the generating series of all evaluations with a fixed underlying graph ( using the combinatorics of the chromatic evaluation of a spin network ) , ( c ) a complete study of the asymptotics of the so - called @xmath1-symbols in all cases : euclidean , plane or minkowskian ( using the theory of borel transform ) , ( d ) an assignement of two numerical invariants ( a spectral radius and a number field ) to spin networks and cubic graphs and ( e ) an explicit illustration of our results for some euclidean , plane and minkowskian @xmath1-symbols ( using the constructive wilf - zeilberger method ) .
atmospheric and solar neutrino experiments provide rather compelling evidence for the phenomenon of flavor oscillations . the celebrated up - down - asymmetry of the atmospheric @xmath4 flux measured by super - kamiokande is consistently explained by @xmath5 oscillations @xcite with the mixing parameters that are summarized in table [ tab : osci ] . the k2k long - baseline experiment provides a first laboratory confirmation , albeit in a pure disappearance mode @xcite . the recent results from sno have largely established active - active flavor oscillations as a solution of the solar neutrino problem @xcite . the lma parameters are strongly favored , but the low case may still be viable ( table [ tab : osci ] ) . neutrino mass differences that are small compared to the ev scale seem to be established . llll evidence&channel&@xmath6 [ @xmath7 & @xmath8 + atmospheric&@xmath5&(1.63.9)@xmath9&0.921 + solar + lma & @xmath10 & ( 0.22)@xmath11&0.20.6 + low&@xmath10 & @xmath12 & 0.92 + lsnd&@xmath13&0.210&(0.2@xmath14 + the only spanner in the works of this beautiful interpretation is the persistence of the unconfirmed evidence for flavor transformations from the lsnd experiment @xcite . if interpreted in terms of neutrino oscillations , the mixing parameters from all three sources of evidence are mutually inconsistent . even including a putative sterile neutrino no longer provides a good global fit because all three sources of evidence prefer active - active over active - sterile oscillations @xcite . it is expected that miniboone at fermilab will confirm or refute lsnd within the upcoming two years @xcite . as there is no straightforward global interpretation of all indications for neutrino oscillations i will follow the widespread assumption that something is wrong with the lsnd signature . if it is due to neutrino conversions after all , something fundamentally new is going on in the neutrino sector . in that case much of the current thinking in this field will have to be revised . this is certainly the attitude that dennis sciama would have taken . he eloquently advocated a cosmological scenario of radiatively decaying dark - matter neutrinos which solves several astrophysical problems , but requires several flavors of light sterile neutrinos and anomalously large electromagnetic transition moments @xcite . no doubt he would have challenged the more conservative approach taken here . the final verdict on neutrino masses and mixings is not yet in , let alone on the more exotic possibilities . the phenomenology of the neutrino sector may turn out to be far richer than envisaged in my presentation of a rather minimal scenario . in what follows i will always assume that there are three neutrino mass eigenstates separated by the atmospheric and solar mass differences . in this scenario , a number of obvious questions remain open . the 12 and 23 mixing angles are large , the 13 mixing angle is small , but how small ? are there cp - violating phases in the mixing matrix ? are the neutrino masses of dirac or majorana nature ? is the ordering of the masses `` normal '' with @xmath15 corresponding to the solar and @xmath16 to the atmospheric splitting , or is it inverted ? and finally , what is the overall neutrino mass scale ? are the masses hierarchical with @xmath17 or degenerate with @xmath18 ? i will focus on these last questions and review the implications of neutrino masses in astrophysics and cosmology . traditionally , cosmology has provided the most restrictive limits on neutrino masses , and this is again the case using large - scale galaxy redshift surveys in conjunction with the standard theory of structure formation . conversely , if the solar lma solution is indeed correct , the cosmic neutrino number density is well constrained by big - bang nucleosynthesis so that a laboratory measurement of the absolute neutrino masses , for example in the katrin tritium experiment @xcite , would directly establish the cosmic neutrino mass fraction . moreover , neutrino masses can have a number of other interesting implications in astroparticle physics in the context of cosmic - ray physics , cosmological baryogenesis , and sn physics . in sec . 2 this review begins with neutrino dark matter and the latest @xmath0 limits from large - scale redshift surveys . 3 turns to the related question of how many neutrinos there are in the universe and how this issue connects with the solar neutrino problem . sec . 4 deals with @xmath19-burst scenarios for producing the highest - energy cosmic rays , sec . 5 with leptogenesis scenarios for producing the baryon asymmetry of the universe 6 is devoted to the time - of - flight dispersion of supernova neutrinos caused by a non - vanishing @xmath0 . finally , sec . 7 summarizes the status of neutrino masses in astroparticle physics . the cosmic number density of neutrinos and anti - neutrinos per flavor is @xmath20 with @xmath21 the number density of cosmic microwave photons , and assuming that there is no neutrino chemical potential . with @xmath22 this translates into @xmath23 . if neutrinos have masses one finds a cosmic mass fraction @xmath24 where as usual @xmath25 is the hubble constant in units of @xmath26 . the requirement that neutrinos do not overclose the universe then leads to the traditional mass limit @xmath27 , an argument that was first advanced in a classic paper by @xcite . later @xcite speculated that massive neutrinos could actually constitute the dark matter of the universe . however , it soon became clear that neutrinos were not a good dark matter candidate for two reasons . the first argument is based on the limited phase space for neutrinos gravitationally bound to a galaxy @xcite . as a consequence , if massive neutrinos are supposed to be the dark matter in galaxies , they must obey a _ lower _ mass limit of some 30 ev for typical spirals , and even 100200 ev for dwarf galaxies . today the most restrictive laboratory limits on the overall neutrino mass scale arise from the mainz @xcite and troitsk @xcite tritium end - point experiments . the current limit is @xcite @xmath28 this limit applies to all mass eigenstates if we accept that the mass differences are as small as indicated by the atmospheric and solar oscillation interpretation . this limit is so restrictive that neutrinos as galactic dark matter are completely out of the question , even without any further appeal to cosmic structure formation arguments . however , cosmic structure formation does place powerful limits on the neutrino mass . the observed structure in the distribution of galaxies is thought to arise from the gravitational instability of primordial density fluctuations . the small masses of neutrinos imply that they stay relativistic for a long time after their decoupling ( `` hot dark matter '' ) , allowing them to stream freely , thereby erasing the primordial density fluctuations on small scales @xcite . while this effect does not preclude neutrino dark matter , it implies a top - down scenario for structure formation where large structures form first , later fragmenting into smaller ones . it was soon realized that the predicted properties of the large - scale matter distribution did not seem to agree with observations and that neutrino dark matter was apparently ruled out @xcite . today it is widely accepted that the universe has critical density and that its matter inventory sports several nontrivial components . besides some 5% baryonic matter ( most of it dark ) there are some 25% cold dark matter in an unidentified physical form and some 70% of a negative - pressure component ( `` dark energy '' ) . and because neutrinos do have mass , they contribute at least 0.1% of the critical density . this fraction is based on a hierarchical mass scenario with @xmath29 , the smallest value consistent with atmospheric neutrino oscillations . an upper limit on the neutrino dark matter fraction can be derived from the measured power spectrum @xmath30 of the cosmic matter distribution . neutrino free streaming suppresses the small - scale structure by an approximate amount @xcite @xmath31 where @xmath32 is the cosmic mass fraction in matter , i.e.excluding the dark energy . this effect is illustrated in fig . [ fig:1 ] where @xmath30 measured by the 2df galaxy redshift survey is shown and compared with the predictions for a cold dark matter cosmology with neutrino fractions @xmath33 , 0.01 , and 0.05 , respectively @xcite . when the theoretical curves are normalized to the power at large scales , neutrinos indeed suppress @xmath30 at large @xmath34 . based on the 2dfgrs data , @xcite and @xcite find limits on @xmath35 in the range 1.83.0 ev , depending on the assumed priors for other cosmological parameters , notably the hubble constant , the overall matter fraction @xmath32 , and the tilt of the spectrum of primordial density fluctuations . for a reasonable set of priors one may adopt @xmath36 at a statistical confidence level of 95% . this limit corresponds approximately to the dot - dashed ( @xmath37 ) curve in fig . [ fig:1 ] , i.e. neutrinos may still contribute as much as 5% of the critical density , about as much as baryons . within the framework of the standard theory of structure formation , the largest systematic uncertainty comes from the unknown biasing parameter @xmath38 which relates the power spectrum of the galaxy distribution to that of the true underlying matter distribution , @xmath39 . the biasing parameter is one of the quantities which must be taken into account when fitting all large - scale structure data to observations of the galaxy distribution and of the temperature fluctuations of the cosmic microwave background radiation . in future the sloan digital sky survey will have greater sensitivity to the overall shape of @xmath40 on the relevant scales , allowing one to disentangle more reliably the impact of @xmath38 and @xmath3 on @xmath30 . it is foreseen that one can then reach a sensitivity of @xmath41 @xcite . for a degenerate neutrino mass scenario the limit of eq . ( [ eq : cosmicmasslimit ] ) corresponds to a limit on the overall mass scale of @xmath1 , far more restrictive than the laboratory limit eq . ( [ eq : tritiumlimits ] ) . however , the katrin project for improving the tritium endpoint sensitivity is foreseen to reach the 0.3 ev level @xcite , similar to the anticipated sensitivity of future cosmological observations . if both methods yield a positive signature , they will mutually re - enforce each other . if they both find upper limits , again they will be able to cross - check each other s constraints . the laboratory limits or future measurements of @xmath0 and the cosmological limits or future discovery of a hot dark matter component can be related to each other if the cosmic neutrino density @xmath2 is known . however , the cosmic neutrino background can not be measured with foreseeable methods so that one depends on indirect arguments for determining @xmath2 . even if we accept that there are exactly three neutrino flavors as indicated by the @xmath42 decay width and that they were once in thermal equilibrium does not fix @xmath2 . each flavor is characterized by an unknown chemical potential @xmath43 or a degeneracy parameter @xmath44 , the latter being a quantity invariant under cosmic expansion . while the observed baryon - to - photon ratio suggests that the degeneracy parameters of all fermions are very small , for neutrinos this is an assumption and not an established fact . in the presence of a degeneracy parameter @xmath45 the number and energy densities of relativistic neutrinos plus anti - neutrinos in thermal equilibrium are @xmath46\,,\\ \rho_\nu&=&t_\nu^4\,\frac{7\pi^2}{120}\ , \left[1+\frac{30}{7}\left(\frac{\xi_\nu}{\pi}\right)^2 + \frac{15}{7}\left(\frac{\xi_\nu}{\pi}\right)^4\right]\,.\end{aligned}\ ] ] therefore , if chemical potentials are taken to be the only uncertainty of the cosmic neutrino density , @xmath2 can only be larger than the standard value . in this sense the structure formation limits on the hot dark matter fraction provide a conservative limit on the neutrino mass scale @xmath0 . conversely , a laboratory limit on @xmath0 does not limit the hot dark matter fraction while a positive future laboratory measurement of @xmath0 provides only a lower limit on @xmath3 . big - bang nucleosynthesis ( bbn ) is affected by @xmath47 in that a larger neutrino density increases the primordial expansion rate , thereby increasing the neutron - to - proton freeze - out ratio @xmath48 and thus the cosmic helium abundance . therefore , the observed helium abundance provides a limit on @xmath47 which corresponds to some fraction of an effective extra neutrino species . in addition , however , an electron neutrino chemical potential modifies @xmath49 . depending on the sign of @xmath50 this effect can increase or decrease the helium abundance and can compensate for the @xmath47 effect of other flavors @xcite . if @xmath50 is the only chemical potential , bbn provides the limit @xmath51 including the compensation effect , the only upper limit on the radiation density comes from precision measurements of the power spectrum of the temperature fluctuations of the cosmic microwave background radiation and from large - scale structure measurements . a recent analysis yields the allowed regions @xcite @xmath52 in agreement with similar results of @xcite and @xcite . however , the observed neutrino oscillations imply that the individual flavor lepton numbers are not conserved and that in true thermal equilibrium all neutrinos are characterized by one single chemical potential @xmath45 . if flavor equilibrium is achieved before @xmath48 freeze - out the restrictive bbn limit on @xmath50 applies to all flavors , i.e. @xmath53 , implying that the cosmic number density of neutrinos is fixed to within about 1% . in that case the relation between @xmath3 and @xmath0 is uniquely given by the standard expression eq . ( [ eq : oega_nu ] ) . the approach to flavor equilibrium in the early universe by neutrino oscillations and collisions was recently studied by @xcite , @xcite , @xcite , and @xcite . assuming the atmospheric and solar lma solutions , an example for the cosmic flavor evolution is shown in fig . [ fig:2 ] . the detailed treatment is rather complicated and involves a number of subtleties related to the large weak potential caused by the neutrinos themselves as they oscillate . the intriguing phenomenon of synchronized flavor oscillations @xcite plays an important and subtle role . the practical bottom line , however , is rather simple . effective flavor equilibrium before @xmath48 freeze - out is reliably achieved if the solar oscillation parameters are in the favored lma region . in the low region , the result depends sensitively on the value of the small but unknown third mixing angle @xmath54 . in the sma and vac regions , which are now heavily disfavored , equilibrium is not achieved . therefore , establishing lma as the correct solution of the solar neutrino problem amounts in our context to counting the number of cosmological neutrinos and thus to establishing a unique relationship between the neutrino mass scale @xmath0 and the cosmic neutrino density @xmath3 . a final confirmation of lma is expected by the kamland reactor experiment @xcite within the next few months of this writing . the number density and mass of cosmic background neutrinos is also relevant for the propagation of extremely high - energy ( ehe ) neutrinos that may be produced by hitherto unknown astrophysical sources at cosmological distances . assuming that neutrinos with energies in the @xmath55@xmath56 ev range are somewhere injected in the universe , and assuming that the cosmic background neutrinos have masses in the neighborhood of 1 ev , the center - of - momentum energy is in the neighborhood of the @xmath42 boson mass . put another way , the required neutrino energy for the @xmath42-resonance is @xmath57 the subsequent decay of the @xmath42-bosons on average produces two nucleons , 10 @xmath58 mesons which subsequently decay into photons , and 17 @xmath59 mesons which subsequently decay into @xmath60 and neutrinos . these @xmath19-bursts would be a source for cosmic rays at the upper end of the observed spectrum , i.e. with energies at and above @xmath61 @xcite . the motivation for considering this sort of scenario is the difficulty of explaining the observed highest - energy cosmic rays which exceed the greisen - zatsepin - kuzmin ( gzk ) cutoff , i.e. which can not reach us from large distances because the cosmic microwave background renders the universe opaque for protons around and above @xmath61 . no compelling explanation for the super - gzk cosmic rays exists , leaving much room and motivation for speculations and exotic scenarios @xcite . of course , the @xmath19-burst scenario is not a real explanation because it appeals to unknown ehe neutrino sources producing huge fluxes and with rather exotic properties @xcite . however , one should view this scenario as an opportunity . the required ehe neutrino fluxes are below current limits , but are accessible to future experiments such as the icecube neutrino telescope at the south pole or the auger air shower array in argentina ( fig . [ fig:3 ] ) . therefore , if such fluxes are detected they provide a handle on sub - ev masses of the cosmic background neutrinos by virtue of the @xmath19-burst production of super - gzk cosmic rays . this would be the only evidence for the cosmic neutrino background and its properties other than provided by big - bang nucleosynthesis . neutrino masses in the sub - ev range can play an interesting albeit indirect role for creating the baryon asymmetry of the universe ( bau ) in the framework of leptogenesis scenarios @xcite . the main ingredients are those of the usual see - saw scenario for small neutrino masses . restricting ourselves to a single family , the relevant parameters are the heavy majorana mass @xmath62 of the ordinary neutrino s right - handed partner and a yukawa coupling @xmath63 between the neutrinos and the higgs field @xmath64 . the observed neutrino then has a majorana mass @xmath65 that can be very small if @xmath62 is large , even if the yukawa coupling @xmath63 is comparable to that for other fermions . here , @xmath66 is the vacuum expectation value of the higgs field which also gives masses to the other fermions . the heavy majorana neutrinos will be in thermal equilibrium in the early universe . when the temperature falls below their mass , their density is boltzmann suppressed . however , if at that time they are no longer in thermal equilibrium , their abundance will exceed the equilibrium distribution . the subsequent out - of - equilibrium decays can lead to the net generation of lepton number . cp - violating decays are possible by the usual interference of tree - level with one - loop diagrams with suitably adjusted phases of the various couplings . the generated lepton number excess will be re - processed by standard - model sphaleron effects which respect @xmath67 but violate @xmath68 . it is straightforward to generate the observed bau by this mechanism . the requirement that the heavy majorana neutrinos freeze out before they get boltzmann suppressed implies an upper limit on the combination of parameters @xmath69 that also appears in the see - saw formula for @xmath0 . the out - of - equilibrium condition thus implies an upper limit on @xmath0 . detailed scenarios for generic neutrino mass and mixing schemes have been worked out , see @xcite for a recent review and citations of the large body of pertinent literature . the bottom line is that neutrino mass and mixing schemes suggested by the atmospheric and solar oscillation data are nicely consistent with plausible leptogenesis scenarios . of course , it is an open question of how one would go about to verify or falsify leptogenesis as the correct baryogenesis scenario . still , it is intriguing that massive neutrinos may have a lot more to do with the baryons than with the dark matter of the universe ! in principle , neutrino masses can be measured by the dispersion of a neutrino burst from a pulsed astrophysical source , notably a supernova ( sn ) . the time - of - flight delay of massive neutrinos with energy @xmath70 is @xmath71 where @xmath72 is the distance to the source . therefore , if a neutrino burst has the intrinsic duration @xmath73 and the energies are broadly distributed around some typical energy @xmath70 , one is approximately sensitive to masses @xmath74 the measured @xmath75 burst of sn 1987a had @xmath76 , @xmath77 , and @xmath78 , leading to the well - known limit @xmath79 @xcite . in a recent re - analysis @xcite find a somewhat more restrictive limit . either way , these results are only of historical interest because the tritium and cosmological limits are now much more restrictive . the neutrino burst from a future galactic sn could yield more restrictive limits because one would expect up to 8000 events in a detector like super - kamiokande for a typical galactic distance of around 10 kpc . with such a high - statistics signal the relevant time - scale @xmath73 is the fast rise - time of around 100 ms rather than the overall burst duration of several seconds . therefore , one is sensitive to smaller masses than the sn 1987a burst , despite the shorter baseline . from detailed monte - carlo simulations @xcite infers that super - kamiokande would be sensitive to about @xmath80 , almost independently of the exact distance . ( at a larger distance one gains baseline but loses statistics , two effects that cancel for a given detector size . ) conceivably this sensitivity could be improved if a gravitational wave signal could be detected preceding the neutrinos , signifying the instant of the stellar collapse @xcite . in this case one may be sensitive to about 1 ev . it is also conceivable that a sn collapses to a black hole some short time after the original collapse . in this case the neutrino signal would abruptly terminate ( within @xmath81 ) , thereby defining a very short time scale . @xcite found that super - kamiokande would be sensitive to @xmath82 . in the foreseeable future megatonne neutrino detectors may be constructed to search for proton decay and to perform precision long - baseline oscillation measurements with neutrino beams . such detectors would have about 30 times the fiducial volume of super - kamiokande . the exact @xmath0 sensitivity for such an instrument has not been worked out . with a megatonne detector one could measure sn neutrinos throughout the local group of galaxies . from andromeda at a distance of 750 kpc one would get around 50 events . using the overall signal duration for @xmath73 yields a sensitivity in the few ev range . the only conceivable time - of - flight technique that could probe the sub - ev range involves gamma - ray bursts ( grbs ) which have been speculated to be strong neutrino sources . if the neutrino emission shows time structure on the millisecond scale , and assuming a cosmological distance of 1 gpc , one would be sensitive to neutrino masses @xmath83 . therefore , observing millisecond time structure in sub - gev neutrinos from a grb would be sensitive to the sub - ev mass scale @xcite . the compelling detection of flavor oscillations in the solar and atmospheric neutrino data have triggered a new era in neutrino physics . in the laboratory one will proceed with precision experiments aimed at measuring the details of the mixing matrix . future tritium decay experiments may well be able to probe the overall neutrino mass scale down to the 0.3 ev range , but if the absolute masses are smaller , it will be very difficult to measure them , and the overall mass scale may remain the most important unknown quantity in neutrino physics for a long time to come . unfortunately , it is unlikely that astrophysical time - of - flight methods will help much . foreseeable sn neutrino detectors are sensitive to ev masses , but not the sub - ev range . on the bright side this means that the measured neutrino light - curve of a future galactic sn will faithfully represent the source without much modifications by neutrino dispersion . cosmological large - scale structure data at present provide the most restrictive limit on neutrino masses of @xmath84 , corresponding to @xmath1 in a degenerate mass scenario . a rigorous relationship between the cosmic hot dark matter fraction @xmath3 and @xmath0 depends on the cosmic neutrino density @xmath2 . if the solar lma solution is correct , big - bang nucleosynthesis constrains @xmath2 without further assumptions about the neutrino chemical potentials . in the lma case neutrinos reach de - facto flavor equilibrium before the epoch of weak - interaction freeze out . while neutrinos do not play a dominant role for dark matter or structure formation , the mass and mixing schemes suggested by the oscillation experiments are nicely consistent with leptogenesis scenarios for creating the cosmic baryon asymmetry . therefore , massive neutrinos may be closely related to the baryons in the universe , not the dark matter . if extremely - high energy neutrinos will be observed in future , the @xmath19-burst scenario provides a handle on the cosmic background neutrinos and their mass through the observed cosmic rays near the gzk cutoff . the great advance in our knowledge of neutrino properties together with the cosmological precision information has rendered the question of neutrino masses in astrophysics and cosmology far more subtle than it looked only a few years ago . we will not know if the accumulation of new results would have persuaded dennis sciama that nature has used massive neutrinos perhaps in different ways for cosmology than he himself had imagined for so long . either way , the connection between neutrino properties and astroparticle physics remains of fundamental interest . this work was supported , in part , by the deutsche forschungsgemeinschaft under grant no . sfb-375 and by the european science foundation ( esf ) under the network grant no . 86 neutrino astrophysics . abazajian , k.n . , beacom , j.f . , and bell , n.f . , 2002 , `` stringent constraints on cosmological neutrino antineutrino asymmetries from synchronized flavor transformation , '' phys . d 66 , 013008 [ astro - ph/0203442 ] . ahmad , q.r . ( sno collaboration ) , 2002a , `` measurement of day and night neutrino energy spectra at sno and constraints on neutrino mixing parameters , '' phys . 89 , 011302 [ nucl - ex/0204009 ] . ahmad , q.r . ( sno collaboration ) , 2002b , `` direct evidence for neutrino flavor transformation from neutral - current interactions in the sudbury neutrino observatory , '' phys . 89 , 011301 [ nucl - ex/0204008 ] . beacom , j.f . , boyd , r.n . , and mezzacappa , a. , 2000 , `` technique for direct ev - scale measurements of the mu and tau neutrino masses using supernova neutrinos , '' phys . 85 , 3568 [ hep - ph/0006015 ] . beacom , j.f . , boyd , r.n . , and mezzacappa , a. , 2001 , `` black hole formation in core collapse supernovae and time - of - flight measurements of the neutrino masses , '' phys . d 63 , 073011 [ astro - ph/0010398 ] . dolgov , a.d . , hansen , s.h . , pastor , s. , petcov , s.t . , raffelt , g.g . , and semikoz , d.v . , 2002 , `` cosmological bounds on neutrino degeneracy improved by flavor oscillations , '' nucl . phys . b 632 , 363 [ hep - ph/0201287 ] . fargion , d. , mele , b. , and salis , a. , 1999 , `` ultrahigh energy neutrino scattering onto relic light neutrinos in galactic halo as a possible source of highest energy extragalactic cosmic rays , '' astrophys . j. 517 , 725 [ astro - ph/9710029 ] . kneller , j.p . , scherrer , r.j . , steigman , g. , and walker , t.p . , 2001 , `` how does the cosmic microwave background plus big bang nucleosynthesis constrain new physics ? '' , phys . rev . d 64 , 123506 [ astro - ph/0101386 ] . weiler , t.j . , 1999 , `` cosmic ray neutrino annihilation on relic neutrinos revisited : a mechanism for generating air showers above the greisen - zatsepin - kuzmin cut - off , '' astropart . 11 , 303 [ hep - ph/9710431 ] .	the case for small neutrino mass differences from atmospheric and solar neutrino oscillation experiments has become compelling , but leaves the overall neutrino mass scale @xmath0 undetermined . the most restrictive limit of @xmath1 arises from the 2df galaxy redshift survey in conjunction with the standard theory of cosmological structure formation . a relation between the hot dark matter fraction and @xmath0 depends on the cosmic number density @xmath2 of neutrinos . if solar neutrino oscillations indeed correspond to the favored large mixing angle msw solution , then big - bang nucleosynthesis gives us a restrictive limit on all neutrino chemical potentials , removing the previous uncertainty of @xmath2 . therefore , a possible future measurement of @xmath0 will directly establish the cosmic neutrino mass fraction @xmath3 . cosmological neutrinos with sub - ev masses can play an interesting role for producing the highest - energy cosmic rays ( scenario ) . sub - ev masses also relate naturally to leptogenesis scenarios of the cosmic baryon asymmetry . unfortunately , the time - of - flight dispersion of a galactic or local - group supernova neutrino burst is not sensitive in the sub - ev range . elementary particles ; dark matter ; cosmology : theory ; supernovae : general ; _ prepared for the dennis sciama memorial volume of new astronomy reviews ( nar ) , edited by f. melchiorri , j. silk , and y. rephaeli . _
the existence of big bang singularity in the very early universe remains an open issue although most of the mysteries of standard hot big bang cosmological model have been addressed by inflationary epoch @xcite . to overcome this initial singularity , ellis @xmath0 @xmath1 proposed a scenario called emergent universe , which is ever existing without singularity and has an almost static behaviour in the infinite past @xcite . eventually the model evolves into an inflationary phase . recently there have been lot of interest in emergent universe models based on standard as well as modified gravity @xcite . this resurgence of interest is because of cmb observations favouring early inflationary universe @xcite and probably this is the possible mechanism for present acceleration @xcite . historically the credit goes to harrison @xcite for obtaining a model of closed universe with radiation and he showed that asymptotically it approached einstein static universe ( esu ) . then after a long gap ellis and maartens @xcite in recent past were able to formulate closed universe with a minimally coupled scalar field @xmath2 . however , exact analytic solutions were not presented in their work and only asymptotic behaviour agreed with emergent universe . subsequently , mukherjee @xmath0 @xmath1 @xcite obtained solution for starobinsky model with features of an emergent universe . also mukherjee @xmath0 @xmath1 @xcite formulated a general framework for an emergent universe using an adhoc equation of state which has exotic behavior in some cases . very recently the idea of quantum tunnelling has been used to model emergent universe @xcite . in most of the general relativity ( gr ) based models , emergent universe is generally obtained for spatially closed past esu which is not stable . contrary to this in modified gravity , emergent scenario can be obtained in spatially flat as well as open / closed model of universe @xcite . also in the context of modified gravity , including braneworld models it has been found , a stable esu can be obtained @xcite . it may be noted that a stable esu also exists in an open universe in @xmath3 gravity , loop quantum cosmology and horova lifshitz gravity @xcite . since astronomical observations including very recent planck results favour spatially flat universe it is imperative to examine emergent scenario in spatially flat universe . in this paper , we look for a possibility of emergent scenario from particle creation mechanism in spatially flat braneworld models . the modern era of brane cosmology began in the context of large extra dimensions , made possible by the hypothesis that the standard model of particle physics is localized on a d - brane . braneworld models of the universe have been the focus of attention in the current decade for the search of many outstanding problems in cosmology . these kind of models are inspired by string theory . in braneworld scenario , our four dimensional universe ( a brane ) is a hypersurface embedded in higher dimensional bulk spacetime . in this brane - bulk scenario , all matter and gauge interactions ( described by open strings ) are localised on a brane while gravity ( described by closed strings ) may propagate into whole space time . this means that gravity is fundamentally a higher dimensional interaction and we only see the effective 4d theory on brane . among the different proposals for brane models , the two prominent are rs ( proposed by randall and sundrum ) and dgp(proposed by dvali , gabadaze and porrati ) models . in rs models , the hierarchy problem could be solved by a warped or curved extra dimension showing that fundamental scale could be brought down from the planck scale to 100 gev . while @xmath4 model have only _ one _ @xmath5-brane @xcite , the @xmath6 model have two _ two _ @xmath5-branes at the ends of the orbifold @xmath7 @xcite . due to its simple and rich conceptual base , rs2 model is very popular and got much attention @xcite and in particular , on the inflationary scenario . the main idea of the dgp model is the inclusion of a four dimensional ricci - scalar into the action .on the 4-dimensional brane the action of gravity is proportional to @xmath8 whereas in the bulk it is proportional to the corresponding quantity in 5-dimensions . the model is then characterized by a cross over length scale @xmath9 such that gravity is 4-dimensional theory at scales @xmath10 where matter behaves as pressure less dust but gravity _ leaks out _ into the bulk at scales @xmath11 and matter approaches the behaviour of a cosmological constant @xcite . in the standard cosmology rs2 model modifies the early universe where as dgp gravity modifies the late universe . so in the present universe brane corrections are not effective in rs2 model . but as phantom energy increases with the expansion , it brings drastic changes in the rs2 model @xcite . this interesting and recent feature of rs2 model which can modify late time cosmic expansion ( if the energy density of the matter content increases at late time e.g. , phantom scalar field ) has also been beautifully shown by using dynamical system tools in ref @xcite . but this kind of modification does not appear if the energy density of the matter content decreases with cosmic expansion e.g , quintessence scalar field , radiation , dust etc . thus rs brane model can also modify late time cosmic expansion in addition to its appreciable impact on early universe cosmology . very recently chakraborty in ref @xcite formulated a emergent universe model in gr using the mechanism of particle creation . the aim of this paper is to formulate an emergent scenario in braneworld as consequence a of particle mechanism . in a sense this paper is a generalisation of the one reported in @xcite to include higher dimensional behaviour . often using bg one obtains cosmological surprises . it may be noted that from thermodynamical aspect it has been proposed that entropy consideration favors the einstein static phase as the initial state of our universe @xcite . the paper is organized as follows : section 2 deals with non - equilibrium thermodynamics from the perspective of particle creation in cosmology while in section 3 we present emergent scenario in braneworld models . the paper ends with a short discussion and concluding remarks 4 . suppose @xmath12 be the internal energy of a closed thermodynamical system having @xmath13 particles . the first law of thermodynamics which is essentially the conservation of internal energy is given by @xcite @xmath14 where as usual @xmath15 is thermodynamic pressure , @xmath16 any co - moving volume and @xmath17 represents the heat received by the system in time @xmath18 . from the above conservation equation , we get the gibb s equation given by @xmath19 @xmath20 , @xmath21 and @xmath22 being the particle number density , heat per particle and energy density of the system respectively . it may be noted that this equation is also true when the particle number is not conserved i.e. , the system is not a closed system @xcite . here we shall consider an open thermodynamical system where the number of fluid particles is not conserved and the non conservation of fluid particles is expressed as ( @xmath23 ) @xmath24 where @xmath25 is the particle flow vector , @xmath26 is the particle four velocity , @xmath27 is the fluid expansion , @xmath28 stands for the rate of change of the number of particles ( @xmath29 ) in a co - moving volume @xmath30 and by notation @xmath31 . the positivity of @xmath28 indicates creation of particles while there is annihilation of particles for negative @xmath28 . any non zero @xmath28 will behave as an effective bulk pressure of thermodynamical fluid and one can use non equilibrium thermodynamics @xcite . in this work , we shall consider a flat , homogeneous and isotropic model of the universe , given by the friedmann - lemaitre - robertson - walker ( flrw ) metric in comoving coordinates ( @xmath32 , @xmath33 , @xmath34 , @xmath2 ) as @xmath35,\ ] ] where @xmath36 is the scale factor of the universe . here we consider universe as an open thermodynamical system which is non equilibrium in nature due to particle creation . + + the energy momentum tensor of the cosmic fluid is characterized by the energy momentum tensor @xmath37 the energy conservation relation of @xmath38 takes the following form @xmath39 where @xmath40 is the hubble parameter . here the pressure term @xmath41 is related to some dissipative phenomena ( say bulk viscosity ) . in the present context , however the cosmic fluid may be considered as perfect fluid where dissipative term @xmath41 is the effective bulk pressure due to particle creation or equivalently the conventional dissipative fluid is not taken as cosmic substratum , rather a perfect fluid with varying particle number is considered . for an isentropic ( or adiabatic ) particle production this equivalence can be nicely described as follows @xcite : using the conservation eqs . ( [ eqthree ] ) and ( [ eqsix ] ) the entropy variation can be obtained from gibb s equation ( 2 ) as @xmath42 where @xmath43 is the temperature of the fluid . but due to isentropic ( or adiabatic ) process , the equilibrium entropy per particle does not change ( as it does in dissipative process ) , i.e , @xmath44 , and from eq.([eq7 ] ) the effective bulk pressure is determined by particle creation rate as @xmath45 hence , the bulk viscous pressure is entirely determined by particle production rate . so we may say that a dissipative fluid is equivalent to a perfect fluid having varying particle number . it may be noted that , although @xmath44 , still there is entropy production due to enlargement of the phase space of the system . in the present context the same effect is due to the expansion of the universe . this effective bulk pressure does not correspond to conventional non equilibrium phase , rather a state having equilibrium properties as well . we note that it is not the equilibrium era with @xmath46 from eq.([eq8 ] ) , we see that if we know the hubble parameter @xmath47 , then the particle creation rate can be determined . further using friedman equations one can relate @xmath28 to the evolution of universe . in what follows as in @xcite , assuming the particle creation rate as a function of the hubble parameter , we can study the cosmological evolution in braneworld . in a homogeneous and isotropic model of universe given by frlw metric ( [ eq4 ] ) , the modified friedmann equation of rs2 model is given by @xmath48,\ ] ] where @xmath49 and the sign of @xmath50 are related to positive and negative brane tension respectively.(for simplicity we are using @xmath51 ) in the absence of bulk pressure , the matter content sector satisfy @xmath52 in order to highlight some characteristics of this model , we review our paper @xcite for the case @xmath53 with @xmath54 , @xmath55 and @xmath56 . using eqs . ( [ eq9 ] ) and ( [ eq10 ] ) , we see that exact solution of energy density is given by @xmath57^{-1}\ ] ] where @xmath58 is the current energy density of dark energy and @xmath59 is the present time . from above equation we see that the braneworld model with @xmath60 @xmath61 that describes an accelerated expansion ends in big - rip singularity in time @xmath62\ ] ] as obtained in most of the gr based models . brane gravity corrections make changes in the behaviour of phantom fluid in case phantom fluid dominates rs2 model of the universe with @xmath53 @xmath63 . in this case it is found that it explains the present cosmic acceleration and exhibits deceleration after a finite time , when energy density grows sufficiently @xcite . in the present work we take @xmath64 , the eq . ( [ eq9 ] ) can be rewritten by introduction of dimensional variable @xmath65 @xmath66 where the effective density is given by @xmath67 it is worth noting that when @xmath68 , the gr based cosmology is recovered and brane effects give new kind of behaviour when @xmath69 . acceleration and deceleration of this model without bulk pressure have been studied in @xcite . the case of @xmath53 is ( also known as bouncing universe ) have been extensively studied in @xcite . also the effect of bulk brane viscosity in this model in the framework of eckart s theory has been studied in @xcite . taking derivative of eq . ( [ eq13 ] ) w.r.t cosmic time @xmath32 and using ( [ eqsix ] ) , we get @xmath70\ ] ] where effective pressures are given by @xmath71\ ] ] @xmath72 and the effective equation of state is given by @xmath73\ ] ] adding eqs . ( [ eq14 ] ) , ( [ eq16 ] ) and ( [ eq17 ] ) and using ( [ eq8 ] ) , we can rewrite eq ( [ eq15 ] ) as @xmath74 again rewriting modified friedmann equation ( 13 ) in terms of @xmath75 , we get @xmath76~~\textrm{where}~~b_{\textrm{rs}}(\epsilon , h)=\sqrt{\frac{6\epsilon h^2}{\lambda}+1}\ ] ] substituting the value of @xmath77 in eq . ( 19 ) , we get @xmath78 } \big(1-\frac{\gamma}{3h}\big)\ ] ] as the particle creation rate is a function of hubble parameter @xmath47 @xcite , we choose @xmath79\ ] ] where @xmath80 and @xmath81 are constants . with @xmath47 in rs - ii model , considering @xmath82 and @xmath83.,scaledwidth=50.0% ] thus using eqs . ( [ eq21 ] ) and ( [ eq22 ] ) , we get the differential equation of @xmath47 describing cosmic evolution as @xmath84 so the hubble parameter can be obtained by solving eqs . ( 23 ) , as @xmath85 , @xmath86}\ ] ] where @xmath87 and @xmath59 is the constant of integration . hence the scale factor can be obtained from the above expression for hubble parameter as @xmath88}\ ] ] where @xmath89 the above cosmological solution shows the following asymptotic behavior for @xmath90 : + @xmath91 , @xmath92 as @xmath93 + @xmath94 , @xmath95 for @xmath96 + @xmath97 $ ] , @xmath98 for @xmath99 . the above asymptotic features show that the above cosmological solution ( described by equations ( [ eq24 ] ) and ( [ eq25 ] ) ) describes a scenario of emergent universe . the modified friedmann equation in this case is given by @xmath100 where @xmath101 is the crossover scale which determines the transition from 4d to 5d behavior and @xmath102 . for @xmath103 , we have standard dgp(+ ) model which is self accelerating model without any form of dark energy , and effective @xmath104 is always non phantom . however for @xmath105 , we have dgp(- ) model which does not self accelerate but requires dark energy on the brane . it experiences 5d gravitational modifications to its dynamics which effectively screen dark energy . the above equation can be rewritten as @xmath106 taking derivative of eq . ( [ eq26 ] ) w.r.t cosmic time @xmath32 and using eq . ( [ eqsix ] ) , we get @xmath107 now using eq . ( [ eq8 ] ) in above equation we get @xmath108 where @xmath109 with @xmath47 in dgp model , considering @xmath111 and @xmath112 .,scaledwidth=50.0% ] as before using modified friedmann eqs . ( [ eq29 ] ) and ( [ eq31 ] ) we get the differential equation of @xmath47 describing cosmic evolution of dgp model as @xmath113 this evolution equation is identical in form to that for rs - ii brane gravity ( i.e. eq.([eq23 ] ) ) and hence it is also poosible to have emergent scenario in dgp brane model in the framework of particle creation mechanism . the present work deals with spatially flat braneworld models in non - equilibrium thermodynamics , based on particle creation mechanism . for simplicity , we assume the thermal process to be adiabatic in nature and as a result the particle creation rate is related to the cosmic evolution . in both rs - ii and dgp brane model , cosmic fluid with constant equation of state @xmath114 , @xmath115 ) is chosen . in both the brane models the particle creation rate is chosen as different functions of the hubble parameter and the brane effect is incorporated through the functions @xmath116 and @xmath117 respectively . however the cosmic evolution equation is similar in both the models and we have emergent scenario in the early phase of the evolution . further it should be noted that the present model of emergent scenario due to particle creation is on ever expanding model of the universe . as the particle creation rate is not constant , so the present model has the basic difference with the steady state theory of fred hoyle @xmath0 @xmath1 @xcite in contrast to the emergent scenario in einstein gravity with constant particle creation rate @xcite . moreover , one can easily check that if we remove the brane effect ( i.e. @xmath118 in rs2 or @xmath119 in dgp model ) then the particle creation rate may be taken to be constant ( by choosing @xmath120 in rs2 model and @xmath121 in dgp brane model ) . the variation of @xmath28 with respect to the hubble parameter has been presented in figs . ( [ fig:1 ] ) and ( [ fig:2 ] ) for rs2 and dgp brane model respectively for various choices of the parameters involved . finally , we conclude that it is possible to have emergence scenario in brane gravity models for non - equilibrium thermodynamics in the framework of particle creation mechanism . ade , p. a. r._et al . _ [ planck collaboration ] , astron . astrophys . * 571 * , a22 ( 2014 ) [ arxiv:1303.5082 [ astro-ph.co ] ] . labrana , p. arxiv:1312.6877 [ astro-ph.co ] . perlmutter , s. j. @xmath0 @xmath122(1999 ) , astrophys . j. * 517,565 ; ( 1998 ) astro - ph/9812133 . canonico , r. and parisi , l. : phys . d 82 * , 064005 ( 2010 ) [ arxiv:1005.3673 [ gr - qc ] ] . wu , p. and yu , h. : phys . b * 703 * , 223 ( 2011 ) [ arxiv:1108.5908 [ gr - qc ] ] . parisi , l. , radicella , n. and vilasi , g. : phys . d * 86 * , 024035 ( 2012 ) [ arxiv:1207.3922 [ gr - qc ] ] . garcia - salcedo , r. , gonzalez , t. , moreno , c. , and quiros , i. : class . * 28 * , 105017 ( 2011 ) [ arxiv:1006.1122 [ gr - qc ] ] . gibbons , g. w. : nucl . b * 292 * , 784 ( 1987 ) . gibbons , g. w. : nucl . b * 310 * , 636 ( 1988 ) . prigogine , i. , geheniau , j. , gunzig , e. and nardone , p. : gen . * 21 * , 767 ( 1989 ) .	here we formulate scenario of emergent universe from particle creation mechanism in spatially flat braneworld models . we consider an isotropic and homogeneous universe in braneworld cosmology and universe is considered as a non - equilibrium thermodynamical system with dissipation due to particle creation mechanism . assuming the particle creation rate as a function of the hubble parameter , we formulate emergent scenario in rs2 and dgp models of braneworld . keywords : emergent universe , modified gravity , braneworld , particle creation , non - equilibrium thermodynamics + + pacs numbers : 98.80.cq , 98.80.-k
this paper is devoted to the understanding of the set of celestial ir ( infrared ) emission bands in the range 3 - 15 @xmath0 m which have been loosely , but wittingly , dubbed unidentified infrared bands ( uibs ) well after their discovery in the 70s . since then they have been thoroughly documented , using higher sensitivity and resolution . particularly large amounts of high quality data were collected by the ir satellites iso [ 1 ] and irts [ 2 ] , which were launched and operated in the 90s . knowledgeable reviews of these and other observations may be found , for instance , in [ 3,4,5 ] . while electronic transitions in atomic ions , such as argon and sulphur , are also detected in the relevant range , sometimes on top of a uib , they are easily distinguished by their much narrower width . the only present consensus regarding the uibs is that they are due to fluctuations of the electric dipole moment associated with the vibrations of the atoms of carbon - rich carriers . different research groups do not agree as to whether the carriers are _ free - flying molecules _ or _ solid - state grains . _ very closely related to this dichotomy is the issue of the excitation mechanism which gives rise to the ir emission . roughly speaking , the free - flying molecule model has been associated from inception to _ stochastic heating _ , i.e. random absorption of a single uv photon , followed by _ immediate thermalization _ of the deposited energy , and subsequent cooling due to ir radiation ; this of course requires very small carriers ( @xmath150 atoms ) and strong uv irradiation . moreover , these carriers are considered to be pahs ( polycyclic aromatic hydrocarbons ) , which are planar clusters of hexagonal c rings with peripheral attached h atoms . no assignment to any known particular pah has yet been proposed , but it is suggested that the problem will ultimately be solved by a size - distribution of neutral / ionized , more or less hydrogenated pahs . in this model , separate particles , the very small grains ( vsg ) , have to be invoked in order to account for the observed underlying continuum , but no composition or structure has yet been specified for this family as well . donn et al . [ 6 ] have discussed this model long ago , but sellgren , tokunaga , boulanger and others [ 5,3,4 ] have recently spelled out some of the `` puzzles '' [ 5 ] presented by the new observations . in this paper , i explore the extent to which these puzzles can be solved by a solid - state model [ 7 ] , and more specifically the _ coal / kerogen _ model . this differs from the pah model on nearly every count . the composition , here , is not restricted to c and h , but also includes small amounts of o , n and s , as suggested by cosmic abundances . _ the structure is not planar or regular , but 3-d and highly disordered ; it is partly aromatic , partly aliphatic and olefinic _ depending on dust environment and history . to accomodate this diversity and disorder , the carrier size should be larger than a few hundred atoms . most importantly , _ well defined natural terrestrial analog materials are proposed _ : these are all members of the kerogen / coal `` saga '' , which have been thoroughly documented during the past half century , and whose composition and structure have been abundantly described in the literature . the diversity of the celestial ir spectra , in relative feature intensities and in position and width , is matched by that of the terrestrial spectra , which itself is not erratic but known to be due to progressive release of s , n , o and h atoms upon ageing and/or heating of the material . no ionization or size distribution need be invoked . the structure accounts for the feature bandwiths and ensures the presence of a continuum where required . finally , the excitation mechanisms invoked here are 1 ) steady state `` thermal '' heating for strongly illuminated dust , 2 ) chemiluminescence for weakly illuminated dust . the coal model was first presented in 1989 [ 8 ] . at that time , modelling the available celestial spectra required the use of relatively highly aromatic coal analogs [ 9 ] . the data now available warrant the complementary use of the lesser aromatic analogs , kerogens [ 10 ] . these data also prompted the consideration and elaboration of a new excitation process [ 11 ] . i therefore believe that the coal / kerogen model is now in a position to cure the main riddles of the pah model . these developments motivated the writing of the present report . section 2 is a short reminder of the main properties and behaviours of kerogens and coals , relevant to the present discussion . in sec . 3 , i delineate some of the main riddles signalled by authors who applied the pah model to the interpretation of astronomical observations . in each case , i then discuss how the solid - state model can remedy the problem or provide an alternative interpretation . section 4 discusses in more detail chemiluminescence as an alternative to stochastic heating . operational conclusions are drawn in the last section . the following is a cursory summary of the properties of natural amorphous carbons of interest to the present issue . details and abundant references to specialized literature can be found in [ 7,10 ] . 2.1 _ composition and structure _ the term coal refers here to the residual part of the raw terrestrial material after minerals and free molecules preexisting in its pores have been disposed of by suitable , standard , physical and chemical procedures . _ this residue is almost exclusively composed of c , h and o , with traces of n and s. _ while coal is extracted from deep in the earth , kerogen is obtained from the superficial sedimentary rocks after treatment by aqueous alkaline and other organic solvents . the solid residue can be considered as coal in dispersed " form ( very small particles ) . _ it is in this form that most of the organic matter in meteoritic carbonaceous chondrites is found _ , a useful hint for our present purposes . the properties of coal do not depend heavily on the geographical origin of the material . as a result , each sample is concisely characterized by one point in the van krevelen diagram ( fig . 1 ) with coordinates o / c and h / c ( atomic ratios ) . the representative points of coals from different origins form an inverted - l - shaped strip . for coals extracted from various mining depths , but same geographical location , the strip nearly reduces to a line of the same shape ; as the mining depth increases , the representative point follows the line down to the origin . the composition and properties of kerogens are more dispersed at high heteroatom concentrations , but ultimately merge with those of coals at low heteroatom concentrations . this graphical orderliness stems from the chemical properties of carbon . like in pahs , the atoms in coals and kerogens are held together by @xmath2 and @xmath3 covalent _ bonds_. but , while the former include only aromatic c _ sites _ , the latter comprise all three types of sites : sp@xmath4 ( acetylenic ) , sp@xmath5 ( aromatic , olefinic ) and sp@xmath6 ( aliphatic ) . this is a result of the presence of impurities " ( o , n , ... ) which impedes the tentency towards the more thermodynamically stable aromatic structures and leads , instead , to _ disordered materials_. as a consequence , pahs know of only one type of c - h bond , while coals / kerogens admit of several others : c - h ( non - aromatic ) , ch@xmath7 ( olefinic , aliphatic ) and ch@xmath8 . other important consequences are : a diversity of c - c bonds ( single , double ) , the presence of coh and c = o groups ( which all have strong incidence on ir spectra ) , and -o- bridges which provide the structure with some flexibility and allow it to grow in 3-d . together with disorder comes the potentiality for _ evolution_. figure 2 illustrates the displacements brought about , in van krevelen s diagram , by spontaneous or induced losses of co@xmath7 , h@xmath7o and ch@xmath9 . in earth , these losses are stimulated by heat and pressure , which explains the trajectory of the representative point in fig . 1 towards the origin as mining depth increases . the structural rearrangements which necessarily accompany these compositional changes are schematically outlined , in fig . 2 , along the left vertical axis , showing the aromatization trend along the evolutionary ( maturation ) track . in the laboratory , mild , prolonged , heating ( annealing ) of a low aromaticity sample prompts its representative point to quickly proceed along the same natural track it would have followed in earth , given time , had it not been extracted . this simple and coherent behaviour of coal / kerogen again emphasizes the tight link between composition and structure and provides a guiding line to find one s way in the _ diversity _ of these materials . thousands of samples have been studied thoroughly all over the world , using all available analytical techniques : thermal , thermogravimetric , ir , raman and nmr spectroscopies , etc . this made it possible to find correlations between spectral bands and assign nearly every band to a _ functional group _ ( a handful of associated atoms ) . these efforts culminated in the building up of representations of the structures of coals and kerogens as a function of their evolutionary stage , i.e. their location in the van krevelen diagram ( fig . 1 ) . the work of behar and vandenbroucke [ 12 ] is a good example of this development . they found that _ it is not possible to adequately describe the observed spontaneous and continuous changes in properties in terms of a limited number of small , specific , molecules ( a conclusion to be extended below to the interpretation of uibs)_. in order to fit the measurements on different samples , one has rather to build a _ random array _ made of a dominant carbon skeleton with functional groups of heteroatoms ( h , o , n , s ) attached randomly to it , then statistically tailor the number of different bondings ( c - c , c = c , etc ) , functional groups ( c - h , c = o , etc ) and aromatic or polyaromatic rings , etc . the variety of environments of the functional groups ensures that their characteristic vibrations will blend into bands of the right width . the relative intensities of the latter will change according to the concentrations of the corresponding functional groups . _ this scheme accounts for the continuous set of ir spectra observed along the evolutionary track_. behar and vandenbroucke give numbers for the structural parameters enumerated above and for 8 representative evolutionary stages . they also give sketches of the corresponding structures , examples of which are shown in fig . a remarkable feature in these sketches is that , even though the size and number of aromatic clusters are quite limited , all the main uibs are present in the corresponding ir spectra , albeit with different intensities . the drawing of the clusters also highlights the 3-d nature of the structures . note the large number of short aliphatic chains in the lesser aromatic samples ; the aromatic clusters become large and dominant in the late stages of maturation . 2.2 _ the ir spectrum _ as a typical example , fig . 4 ( adapted from [ 13 ] ) displays the ir absorbance spectrum of a young kerogen of type ii ( h / c=1.32 , o / c=0.104 ) . the peak wavenumber ( @xmath10 ) and wavelength ( @xmath0 m ) , integrated intensity ( cm / mg ) and assignment of the bands are , respectively , ( see [ 14 ] ) : \1 . @xmath113400(@xmath112.95 ) ; 46.2 ; oh stretch ; due partly to chemically bonded oh groups and partly to adsorbed or trapped , not chemically bonded , h@xmath7o molecules ; the latter are responsible for the long redward tail , through h - bondings with other parts of the skeleton ; peak position and band profile change considerably with h@xmath7o content ( depending on evolutionary stage or heat treatment ) . this feature is clearly distinct from the water ice band , which peaks near 3.1 @xmath0 m and has a much steeper red wing ( associated to that observed in the sky towards young stellar objects ) . adsorbed h@xmath7o is probably absent from is dust . 3060 ( 3.27 ) ; c - h aromatic or olefinic stretch , barely visible here , and only measurable for very small h / c and o / c ratios . \3 . 2920 ( 3.42 ) ; 48.2 ; blend of anti - symmetric and symmetric ch@xmath8 stretch at 2962 and 2872 , asymmetric and symmetric ch2 stretch at 2926 and 2853 ; ch stretch at 2890 @xmath10 , respectively . here , _ the non - aromatic c - h bonds reside mainly on non - aromatic structures , not at the periphery of pahs like in the work of wagner et al . [ 15 ] . _ \4 . 1710 ( 5.85 ) ; 14 ; c = o ( ketone ) stretch . 1630 ( 6.15 ) ; 11.4 ; disputed ( but perhaps concurrent ) assignments to h@xmath7o deformation , quinonic c = o with h bond and c = c olefinic and aromatic stretch . 1455 ( 6.87 ) ; 6.1 ; asymmetric ch@xmath7 and ch@xmath8 deformation . 1375 ( 7.28 ) ; 1.2 ; symmetric ch@xmath8 deformation . 1800 to 900 ( 5.5 to 11 ) ; 67.3 ; massif ( underlying broad band ) peaking near 8 @xmath0 m , due to c ... c and c - o stretch , c - h in - plane bend and oh deformation ; in highly aromatic samples , there may be a contribution by electrons , in the form of a plasma resonance [ 16 ] . \9 . 930 to 700 ( 11 to 14 ) ; aromatic out - of - plane bending , depending on the number of adjacent protons ; weakly contrasted 3 or 4 peaks , barely visible here ; intensity increases with mining depth but never exceeds an intensity of 5 . . an underlying continuum , roughly decreasing with frequency , is mostly visible in the near - ir , but also present farther to the red in all coals / kerogens ; part of it is due to light scattering ( in transmission measurements ) and part of it to real absorption ; it increases strongly with aromaticity . the intensity , k , given here ( third number ) is related to the absorbance , @xmath12 , by @xmath13 where @xmath14 is the material density in g @xmath15 ( @xmath111 ) and @xmath16@xmath17 , the bandwidth in @xmath10 . _ k ( cm / mg ) is @xmath18 times the intensity usually defined in the astrophysical literature _ ; it is also related to the integrated cross - section per c atom , @xmath19 , by @xmath20 not unexpectedly , a strong relationship holds between the bands at 2920 and 1455 @xmath10 for all kerogen types and evolutionary stages : the ratio of their integrated intensities is @xmath118 . also , k(3.4 @xmath0 m ) increases linearly with h / c , from 0 at @xmath10.3 at least up to 80 cm / mg at h / c=1.3 . it should be clear from the above that a single spectrum like fig . 4 can not , by far , convey the full diversity of coal and kerogen spectra . series of spectra in restricted wavelength ranges and corresponding to successive degrees of maturation may help for this purpose . an example is given in fig . 6 for the 3 @xmath0 m range ; others can be found in [ 10 , 14 ] . here i discuss some of the problems encountered in applying the pah model to the most recent astronomical observations . in enunciating these , i shall try to remain faithful to the wording of the referenced papers ; this task is made easier by the authors candour . the recent review by sellgren [ 5 ] was particularly useful in this respect . 3.1 _ carrier composition _ by definition , pahs are pure hydrocarbons . however , the carriers of the uibs are most likely formed in the winds of c - rich post agb stars . in general , oxygen is not much less abundant than carbon in these winds . nitrogen and sulphur are also present in notable quantities . it is therefore difficult to understand why these three elements should be excluded from the dust formed in such environments . all the more so since the less abundant silicon shows up in the sic feature at 11.3 @xmath0 m , and , most importantly since the ism ( interstellar medium ) harbors many families of molecules containing these so - called heteroatoms , especially o , mostly associated with c ( see [ 17,18 ] ) . on the other hand , minority atoms , mostly o , are present in coals and kerogens in known concentrations , and these inclusions have important consequences : 1 ) they considerably enhance the dipole moment of some vibrational transitions involving c atoms , such as the 6.2 @xmath0 m stretch of the c = c bond ; 2 ) they give rise to features observed in the sky in the mid - ir , such as the band at 1710 @xmath10 ( 5.85 @xmath0 m ) and the massif at 7 - 9 @xmath0 m [ 10 ] , as well as in the far ir , such as the 30 @xmath0 m blend associated with coh wagging ( see [ 19 ] ) ; 3 ) minority atoms efficiently interfere with the tendency of c atoms to settle into an ordered aromatic structure ; _ hence the amorphous 3-d structure of fig . 3 , which determines all the favourable properties of this model _ , as shown below . 3.2 _ feature bandwidths _ the uibs are strikingly wider than the bands of any observed and identified molecule , the narrowest width being about 15 @xmath10 . even at the highest resolution , no rotation - vibration decomposition is discernible , although such decomposition was clearly demonstrated , with the same spectrometers , for many identified molecules , such as h@xmath7 and h@xmath7o . neither is it possible to assign these widths to lifetime broadening , since the typical width of 30 @xmath10 would correspond to a lifetime of about 0.1 ps , way shorter than the radiative lifetime ( @xmath110.1 s ) and non - radiative lifetimes , ranging between 1 and 1000 ps . thermal doppler effects are also excluded for the masses and temperatures of particles of interest . thus , although uibs can be fit by one or more lorentzians or gaussians , these are too wide to carry physical meaning : the uibs are definitely broader than laboratory pah features _ in absorption _ ( see [ 20 ] , [ 5 ] ) . in _ emission _ , highly excited molecules may have much broader features than in absorption because of _ anharmonicity and spectral congestion_. this is illustrated by williams and leone ( 21 ] , who excited gas - phase naphtalene molecules ( plane clusters of 2 benzenic rings , 18 atoms ) with short laser pulses of photons of energy 5.3 and 6.6 ev . monitoring the outgoing radiation at 3.3 @xmath0 m , ( c - h stretch ) , they found its width to decrease from @xmath11100 to @xmath1130 @xmath10 , while the peak shifted to the blue by 45 @xmath10 , as the energy content of the molecule was reduced by on - going radiative emission . no such behaviour is observed in the sky . being associated by birth with single - photon excitation ( stochastic heating ) , the pah model is therefore faced with a dilemma : either the molecule is so small , and the photon so energetic , that the `` temperature '' can reach about 1000 k and the emission is strong enough as to be detected , but then with unacceptably large widths and red shifts ; or the molecule is large and the photon energy low so that no undesirable broadening and/or red shift occurs , but then with unacceptably weak intensities and narrow widths . recent attempts to solve this dilemma relied on the strong but unsupported assumption that the band position and widths of the features are the same functions of temperature for all interstellar pah sizes , or on very specific mixtures of selected pah cations supposed to be independent of history and environment ( see discussions by verstraete et al . [ 22 , 23 ] and boulanger et al . [ 24 ] ) . by contrast , the _ disordered solid - state _ model assigns the features to the synchronous vibrations of specific functional groups , and most of their width to the dispersion of similar functional groups among various environments , a situation corresponding to small - scale inhomogeneity [ 20 ] . the widths of uibs is indeed typical of solid , disordered materials [ 25 ] and , in particular , coals and kerogens [ 10 ] . the disorder , here , is mainly induced by heteroatoms and is , therefore , chemical and structural in nature ( sec . 2 ) . within limits , and especially below 500 k , feature widths are not sensitive to temperature or grain size , provided it is large enough ( a few hundred atoms ) to accomodate the required diversity of environments for a given average chemical composition . under such circumstances , small degrees of ionization and dehydrogenation are not likely to alter the spectrum notably , in stark contrast with pah behaviour . while this explains the remarkable stability of the uib widths [ 3,22 ] , it also leaves room for the observed diversity [ 20 ] , which may result from different compositions at inception , as well as different thermal and radiative processing in the interstellar medium . because of this disordered , solid - state structure , the coal / kerogen model also adequately mimicks the celestial massifs at 7 - 9 @xmath0 m ( see fig . 4 ) and 11 - 13 @xmath0 m [ 9 ] , which are major puzzles for the pah model [ 5 ] . 3.3 _ carrier aromaticity _ the use of the term pah implies full aromaticity . this radical assumption was initially suggested by the strength and ubiquity of the 3.3 , 6.2 and 11.3 @xmath0 m features . however , it was not supported by later observations illustrated by fig . 5 for the stretch bands and in [ 7 ] for other ones : distinct emission features between 3.3 and 3.6 @xmath0 m are seen towards agb stars such as iras 05341 + 0852 , 04296 + 3429 and crl 2688 , in planetary nebulae ( e.g. hb 5 [ 26 ] ) and towards at least one nova : nova cen 1986 [ 27 ] . in some cases , they are even stronger than the 3.3 @xmath0 m feature , as in absorption towards the galactic center , where the aromatic feature is dwarfed by a massif with peaks near 3.4 and 3.5 @xmath0 m . the initial assignment of these `` red '' features to hot bands ( v@xmath211 ) of the c - h str transition has proven unconvincing because of their strength and the non - detection of the correlative first harmonic to be expected near 1.67 @xmath0 m [ 15 , 26 ] . it seems therefore more reasonable to assign them to the well documented aliphatic and olefinic , symmetric and asymmetric , c - h stretches at 3.4 , 3.46 , 3.515 and 3.56 @xmath0 m , in ch , ch@xmath7 , ch@xmath8 and aldehydes ( -cho ) groups [ 33,34 ] . figure 6 illustrates this statement by setting , side by side with the spectra of fig . 5 , the spectra of coals at increasingly higher stages of aromatization ( see sec . 2 ) : not only is the fit of corresponding spectra remarkable , but the regular decrease of the red blend together with the rise of the 3.3 feature confirms their assignment to non - aromatic vibrations . neither is it obvious that the 6.2 and 11.3 @xmath0 m features are characteristic of aromatics indeed , while the 11 - 13 @xmath0 m massif with 3 or 4 peaks , observed towards post - agb stars may reasonably be ascribed to the out - of - plane vibrations of aromatic c - h bonds [ 9 ] , it has proven very difficult to devise a pah cluster that could display a highly contrasted , strong and lonely 11.3 feature as observed in most is ( interstellar ) spectra . according to [ 35 ] , this requires `` rectangular '' pahs with 100 to 150 c atoms , a very special requirement indeed ! ( see next subsection ) . moreover , the expected correlation between the 3.3 and 11.3 features is not confirmed by observations [ 47 , 52 , 63 ] . even the exact spatial coincidence of the intensity peak of the 11.3 feature with that of the 3.3 feature appears to be in doubt ( see [ 63 , 64 ] ) . on the other hand , a ( more familiar ) alternative or concurrent assignment for the 11.3 @xmath0 m band could be the wagging of the two h atoms of a vinylidine group ( r@xmath7-c = c - h@xmath7 ) , which also occurs at the desired frequency [ 33 ] . clearly , the constraint of full aromaticity is not warranted : an understanding of the 3.4 " feature , as well as of the diversity of uib spectra , is better served by thinking in terms of _ evolution of is dust composition and structure _ , both intimately connected and induced by prolonged heating ( sec . 2 ) and irradiation [ 36 ] . while this is natural , straightforword and quantitative in the coal / kerogen model , it seems to be hardly possible with compact pahs , whatever their size and degree of ionization or hydrogenation . 3.4 _ spectral continuum _ known pahs , even when ionized , can not provide enough visible _ continuum _ absorption to account for the energy in the emission features observed around uv - poor reflection nebulae [ 5].neither can they mimick the unpolarized near - ir continuum detected towards the same objects [ 5,3 ] . this , of course , does not come as a surprise since free - flying molecules are not usually known for their continuum , although they do exhibit lines or bands in the vis / uv . herein lies yet another , perhaps the main , reasons for invoking pahs much larger than initially envisioned [ 5,35 ] . unfortunately , none has yet been produced in the laboratory , if this is possible at all . indeed it is well known that , in the absence of externally applied pressure , a planar sheet of compact benzene rings ( graphene ) of increasing size tends to bend , warp and finally folds upon itself to form a fullerene or a carbon nanotube[37 ] . now , these particular carbon structures indeed have features in the uv / vis but they also have features elsewhere , such as the four conspicuous ir bands of fullerene at 526 , 576 , 1183 and 1428 @xmath10 [ 38 ] . none of these has been detected in the sky despite intensive seach . moreover , even these structures do not display mid- and far - ir continua , which points to still another puzzle of the pah model : there is also a _ mid - ir _ continuum underneath the uibs ; for want of a better assignment , this has been attributed to a separate dust component , the very small grains ( vsgs ) , but its composition and structure could not be specified further [ 5 , 39 ] . note that they were thought to be small solely on the grounds of the self - imposed constraint that their emission should be excited upon absorption of a single uv / vis photon . recent surveys of the spatial variations of this continuum have added to the confusion and seem to require again very special behaviour of the assumed separate component in order to match the systematic changes in the ratio of features to continuum [ 24 , 39 , 40 , 41 , 42 ] . of course , theorists never lack an explanation , but all tentative applications of the pah model seem to converge towards particles much larger than invoked in the first place [ 5 , 24 , 35 , 43 , 44 ] . at that point , it must be recalled that _ bulk _ carbonaceous materials indeed display a continuum , but the profile and intensity of the latter depend heavily on the composition and structure of the material [ 7 ] . thus , the continuum of glassy carbon is very strong and varies approximately as @xmath22@xmath23 at least as far as 100 @xmath0 m , but coals and a - c : h ( amorphous hydrogenated carbon ) are semi - conductors whose absorption falls down sharply towards the red at a cut - off wavelength ( @xmath111 @xmath0 m ) which increases with aromaticity . the lower aromaticity kerogen has a cut - off at shorter wavelengths and , hence , a weaker mid - ir continuum , better suited to the high contrast of the uibs ; it has also a near - ir continuum rising towards short wavelengths ; it has the further advantage of carrying both continuum and bands [ 10 ] , thus eschewing the need for two different carriers . 3.5 _ near- and far - uv bands _ simple aromatic molecules are known for their near- and far - uv transitions between ground and excited electronic states[45 ] . none of these was detected in the sky , either in absorption or in emission ( fluorescence ) [ 6 ] , although several attempts were made to identify the dibs ( diffuse interstellar bands ) with such bands . in this field , too , hopes now hinge upon large 3-d systems [ 46 ] . indeed , it was shown that high - temperature treatment ( htt ) of coals favored aromatization and the growth of a _ single _ near - uv feature which irreversibly becomes narrower and bluer [ 7 ] . in some favourable cases , the end point of this evolution is _ polycrystalline ( disordered ) graphite _ , with a single , conspicuous , feature near 217 nm , on top of a continuum which increases sharply to the blue and decreases slowly to the red [ 7 ] . then , of course , the material can hardly be designated as a pah . 3.6 _ infrared bands _ regarding wavelength and bandwidth , the uibs are remarkably stable against large variations in the environment and irradiation . this stability is incompatible with an assignment of the bands to a soup of free - flying independent pahs , neutral or ionized and more or less hydrogenated , as suggested by a number of authors in the past , since later laboratory studies showed that individual pahs are highly sensitive to irradiation wavelength and intensity ( see discussions in [ 5 , 24 , 48 , 49 , 50 ] . in trying to use these parameters to fit observations , one is often led to conflicting or paradoxical conclusions ( see discussions in [ 22 , 23,35 , 49 , 50 , 51 ] ) . one is therefore led again to prefer a solid - state ( bulk ) model : in that case , ionization and dehydrogenation can only affect the surface of the grain , a minor fraction of the whole . also , the dust is then much more robust against destruction by shocks . moreover , the structure and composition of a bulk , amorphous , material as coal / kerogen can _ evolve slowly , spontaneously or upon heat treatment _ as shown in sec . this is accompanied by subtle changes in band wavelengths and widths ( small fractions of a micrometer ) , also observed in the sky ( see [ 10 , 20 ] ) . very large changes in relative intensities are also observed , especially in the ratios [ 3.3]/[3.4 ] , [ 3.3]/[11.3 ] and [ 11.3]/[12.7 ] ( see[47 , 52 ] ) . on terrestrial analogs , most of these can be explained by coherent and systematic changes of composition and structure brought about by heat or slow spontaneous evolution . the most spectacular and best documented case is that of c - h stretch ( see fig . 6 ) : dischler analyzed in detail the evolution of each component of this group of vibrations [ 34 ] . thus _ the spectral diversity of coals / kerogens of different degrees of maturation ( ranks ) and origins matches the diversity of astronomical spectra . the main astronomical features associated with carbonacious is dust , including the continuum , are provided , to various extents , by coals / kerogens of various ranks . different coals / kerogens or mixtures thereof can account , roughly at least , for astronomical spectra from different is environments . these models introduce no significant spurious feature anywhere in the explored spectrum . _ examples of quantitative fits between uibs and models can be found in [ 7 ] for the uv / vis range , here in fig . 5 and 6 for the near - ir , in [ 10 , 11 ] for the 5 - 10 @xmath0 m range , in [ 9 ] for the 11 - 13 @xmath0 m range and [ 19 ] for the far - ir range ; see also [ 32 ] . that is not to mean that a perfect match between celestial and laboratory spectra is easily obtained . for one thing , the high quality of iso spectra has revealed many new weak features , or details of the profiles of previously known features , which have yet to be elucidated ( see [ 35 , 52 ] ) . more importantly , no satisfactory interpretation is yet available for the 12.7 feature , nor for the 11.3 feature being usually so much stronger than the underlying blend ( 11 - 13 @xmath0 m ) which is commonly assigned to c - h mono , duo , trio and quarto out - of - plane bends ( see sec . _ the 11.3 and 12.7 bands are not present with a high contrast in the solid- state model spectra . _ but the coal / kerogen model provides at least a framework with definite guidelines and valuable assets : -representative natural samples are available with known composition , structure , physical properties and band assignments ; -well - defined heat treatments can alter these properties controllably and coherently ; -the range of possible model spectra can be extended by using samples of different depths and geographical origins , thus scanning the whole van krevelen diagram ( fig . 1 ) , i.e. a large range of _ natural _ structures . -the existence , in space , of carbon chains and molecules as large as acetone is easier to understand in terms of breakup of large , inhomogeneous , kerogen - like , grains ( formed in dense " circumstellar shells and destroyed in strong is shocks ) , rather than in terms of synthesis in the tenuous is medium . 3.7 _ excitation _ the relatively high ratio of short to long wavelength uib intensities initially led to the conclusion that the emitter temperature should be very high . under the assumption of stochastic heating by absorption of a photon , and for the relatively small pahs initially considered , calculations showed that such a temperature already requires the absorption of uv to far uv photons [ 53 ] . the steady increase of the absolute intensities of the uibs with ambient uv radiation field intensity ( g value ) over 4 to 5 orders of magnitude has been considered as the strongest argument in favour of stochastic heating by uv photons [ 4 ] . these conclusions are in conflict with a number of astronomical and laboratory facts . 1)the relative uib intensities from reflection nebulae are independent of the temperature of the illuminating star , i.e. of the hardness of the radiation field , from 22000 down to about 5000 k ( see [ 5 ] ) ; this is at variance with the sharp decrease of dust temperature with photon energy predicted by calculations [ 53 ] . 2)shortward of 10@xmath0 m , the shape of uib spectra of several galaxies that have been studied are strikingly similar for nuclear regions , active regions in the disks , as well as quiet regions [ 49 ] . a similar stability is observed among spectra of regions of hugely different g values , which are accompanied by large variations of star temperature and , hence , of the spectrum of the exciting radiation . again , one should therefore expect notable non - linear variations of relative band intensities , which is not the case [ 4 ] . the authors of [ 49 ] conclude that the global relation between uib intensity and uv illumination might be only indirect . this is corroborated by the observation of uib emission from poor - uv regions , such as galaxy m31 [ 61 ] . 3)the small pah - stochastic heating " paradigm is also confronted with the stability of the uib wavelengths and widths against environmental variations , especially in view of attendent dehydrogenation , ionization or downright destruction of the molecule [ 4 , 5 , 24 , 39 , 48 , 50 ] ( see sec . 4)it appears from the previous subsections that all available spectral evidence weighs in favour of bulk , solid - state models of is dust , as opposed to free - flying individual molecules . recent discussions involving pahs also invoke large molecules " , up to 1000 atoms or more [ 5 , 24 , 35 , 43 ] . now , according to the former calculations , this leads to prohibitively low dust temperatures and undetectable uib emission . one way out of this paradox is to hypothesize very special , but as yet not measured , electronic properties : high photon absorption cross - sections , absorption cut - off sliding to the red as the particle size increases , etc . this becomes still more accrobatic if one is to account for uibs detected in poor - uv environments ( see [ 22 , 54 ] ) . the coal model is confronted with a symmetrical puzzle as long as it holds to the simplest excitation process : thermal equilibrium . this is only effective in strong radiation fields , in the immediate vicinity of hot stars , as is the case with post - agb stars and ppne ( pre - planetary nebulae ) ; see [ 9 ] ) ; but it is out of the question in rne ( reflection nebulae ) and questionable in other instances . all models are therefore confronted with the task of conceiving a third excitation mechanism . a significant hint towards this goal is provided by the spatial distribution of the intensity of the 3.3 feature across the orion bar observed edge - on [ 55 ] : it forms a narrow peak between the ionization front ( delineated by the br @xmath12 and p@xmath12 lines ) and the h@xmath7 excitation peak ( delineated by its ro - vibration transitions in the ground state ) . while the sharp drop towards the ionization front may perhaps be attributed to partial or total destruction of the carrier in the ( hot ) hii region , the drop on the opposite side is more problematic because it can not be due to a decrease in the column density . sellgren et al . [ 55 ] finally assigned it to the extinction of the stellar radiation by molecular hydrogen , assuming a density of 10@xmath24 @xmath15 for the latter . however , if this was the case ( with a small pah , as required by single - photon heating ) , one would also expect a change in the spectrum of the exciting radiation and , consequently , a change in the 3.3 emission feature across the bar . what is observed , in fact , is a remarkable stability of wavelength , width , [ 3.4]/[3.3 ] and [ continuum]/[3.3 ] ratios . now , the steep , large and opposite intensity variations of the ionic and molecular hydrogen spectral features observed to occur across the bar suggest that , in between the ionic and molecular regions , their must be a narrow transition layer of atomic hydrogen , roughly coincident with the 3.3 feature peak . a similar situation has been described in ngc 7027 [ 56 ] and seems to have been observed at the edge of high - latitude cirruses [ 41 , 65 ] . this led us to surmise that the excitation of the uibs is linked with the presence of these atoms , rather than directly with the ionizing radiation [ 57 ] . such a scenario ( developed in the next section ) naturally and straightforwardly explains the small transversal extent of the 3.3 layer , without having recourse to uncertain or qualitative assumptions as to the destruction of dust carriers towards the exciting star [ 41 ] , or steep extinction of the uv towards the molecular cloud . hydrogen _ radicals _ ( free atoms ) are known to be highly reactive , especially with hydrogen and carbon . when such a radical hits a hydrogeneted surface , it is likely to form , with an h atom of the surface , a h@xmath7 molecule , which will promptly escape , leaving a _ dangling bond_. this catalytic reaction between two h atoms is the main route envisioned by astrochemists to explain the high rate of formation of molecular hydrogen in space : here , the solid surface provides the third body which is necessary to ensure conservation of momentum . since carbon is known for its high catalytic efficiency , our is carbon dust is a good candidate for the job , provided it can present a large fraction of dangling bonds per c atom . it must therefore be disordered and porous , as are coals and kerogens . note that the nascent h@xmath7 molecule carries away a large part of the energy made available by this exothermic reaction ; it must therefore be ro - vibrationally excited and this may be the source of part of the h@xmath7@xmath25 lines observed in pdr s and at the edge of molecular clouds , between ionized h and molecular gas ( co ) . now , when another h atom approaches the dangling bond left on the dust , it is very likely to be captured and form a new strong c - h bond . the chemical ( potential ) energy of this bond , @xmath114.5 ev , is insufficient to dissociate other bonds or excite higher electronic states ; it is then available as ground - state vibrational energy for the atoms in the grain . initially , this energy is essentially in the form of c - h stretch . however , it quickly spreads around , exciting the other characteristic ( normal ) modes of the given particle . the bond anharmonicity ( non - linearity of restoring forces ) opens channels for the available energy to be shared among the modes , which otherwise would be isolated . this process is called ivr ( internal vibrational redistribution ) , and is usually very short - lived , of order 1 ps , after which the energy deposited in an isolated grain can only escape through ir radiation , with a time constant of order 0.1 s. such _ infrared chemiluminescence _ , following molecular reaction in the gas phase was observed long ago [ 58 ] and often thereafter . similar processes occur in an isolated gaseous molecule upon absorption of a vis / uv photon , as in the experiment of williams and leone [ 21 ] , referred to above . the problem of theoretically predicting the chemiluminescence spectrum is extremely difficult to solve because of lack of detailed knowledge of the various couplings ( force fields ) between atoms in a given molecule : electrostatic dipole and higher order moments , van der waals and other forces , hydrogen bonding , etc.this is compounded by the length of dynamical calculations in the case of large systems , such as those that are now envisioned as dust models . fortunately , commercial software is now available which is dedicated to numerical simulation of molecular structure , dynamics and reactions . i have used such gear to collect quantitative , semi - quantitative and qualitative data on the structure of coal / kerogen model particles as large as 500 atoms , the interaction of h atoms with such particles and the resulting vibrational spectra [ 59 , 60 ] . the main result of these computations , of interest for present purposes , is that , to a large extent , _ the predicted emission spectrum is very similar to the absorption spectrum , independent of particle size and colour temperature of the illuminating star . no immediate thermalization occurs_. moreover , since the only relaxation route is through ir radiation , _ the luminescence efficiency in the uibs should be very high _ ( of order 0.1 if it is defined as the ratio of ir energy emitted over deposited energy ) . chemiluminescence thus solves the puzzles encountered above with the two other excitation mechanisms , _ provided enough atomic h is available_. if there is a luminous star near by a molecular cloud , it will give rise , at the edge of the cloud , to a pdr ( photodissociation r egion ) such as the orion bar ( referred to in sec . 3.7 ) , with contiguous ionized , atomic and molecular regions , successively in the radial direction from the star . quantity _ of chemically excited dust in the atomic layer , and hence the uib intensities , will obviously depend on g , but the ir emission spectrum will not ! this is all compatible with the roughly linear increase of the individual uib intensities versus g , as well as with the stability of their spectrum over 5 orders of magnitude of g [ 24 ] . moreover , because the rates of dissociation of h@xmath7 molecules and ionization of h atoms depend on the flux of uv photons , the uib intensities will depend , in absolute but not relative value , on the stellar colour temperature , which is not the same for all g s . this , together with the random variations of molecular density at the cloud edge , accounts for the observed excursions from strict linearity of intensity with g [ 24 ] . of course , this does not exclude the coexistence of thermal equilibrium heating of grains if the ambient radiation field is strong enough . in the opposite case of no neighbouring luminous star , atomic h is still present , but in smaller quantities , if only for the diffuse galactic is radiation field ; this accounts for the uib luminosity of nebulae at high galactic latitudes , such as the chameleon cloud or the ursa major cirrus [ 41 ] . a further observational support for this excitation model is the spatial coincidence of the uib emission with the filament " of excited h@xmath7 emission at the edges of ngc 7027 [ 56 ] and of the @xmath14oph main cloud [ 65 ] : indeed , as indicated at the beginning of this section , chemiluminescence induced by atomic hydrogen is genetically associated with the production of hydrogen molecules which carry away part of the chemical energy of the incoming h atom and are , therefore , ro - vibrationally excited . in application of these ideas , a chemical kinetic model was built [ 11 ] for the galaxy m31 , whose diffise ism emits the uibs although it is nearly devoid of far uv radiation [ 61 ] . a good numerical fit to the observed ir intensities was obtained assuming an atomic h density between 50 and 200 @xmath15 ( in agreement with the range determined for the edge of a typical molecular cloud in our galaxy ) and a chemiluminescence efficiency of 0.3 , in agreement with the numerical chemical simulation referred to above . the essential difference between this excitation mechanism and stochastic heating does not lie in the energy deposition process ( atomic h capture or photon absorption ) , but in the unwarranted assumption at the core of the stochastic heating theory , that _ thermalization of the absorbed energy occurs before radiative relaxation _ , only because the latter may require up to 1 s. common wisdom mixes up ivr with thermalization and the last decades of the past century were necessary to discriminate between various types of _ chaos _ and between various degrees of _ randomization _ ( see bibliography in [ 60 ] ) . in momentum - configuration phase space , thermal ( statistical ) equilibrium is the extreme case where all points of the energetically available phase space are equally probable and ultimately visited by the representative point of the molecular system . this stage is only ideal and , in fact , very difficult to reach on a microscopic scale and in the finite time of experiments , but it is easier to approach when dealing with macroscopic assemblies of identical particles interacting with each other and with a thermal bath " ( itself an ideal concept ! ) . by contrast , experiments resolved in time , space and wavelength have shown that the vibrations of atoms in an isolated system may retain a high degree of coherence for a long time . _ that means that only parts of phase space are visited , that randomization is not complete and that a temperature can not be defined_. i have shown numerically that this is the case for a disordered system , even as large as 500 atoms [ 60 ] . for this purpose , i used a structure inspired by chemical models of coal [ 62 ] . a small fragment of this is shown in fig . 7 , and fig . 8 displays the frequency spectrum of its elecric dipole fluctuations after deposition of @xmath114.5 ev in one c - h bond . if the same energy were thermalized , the peak temperature reached by the particle would be @xmath11 300k and the ratio [ 3.3]/[11.3 ] would fall below 1@xmath26 ! experiments such as that of williams and leone [ 21 ] and wagner et al . [ 15 ] seem to show that randomization in pahs is also slow after absorption of a uv photon , so that ir _ _ photo__luminescence is observed . this process could therefore coexist with chemiluminescence . however , it probably suffers from the undesirable loss of part of the deposited energy to vis / uv fluorescence and electronic continuum emission ; neither of these affects chemiluminescence since the chemical excitation considered here leaves the system in the electronic ground state . one is overwhelmed by the immense diversity of both uib and coal / kerogen ir spectra . some relief is provided by 1 ) the sparsness of the prominent bands , 2 ) their relative stability in wavelength , 3 ) the encouraging fits attempted between observations and models over most of the uibs and extended to the uv / vis and far ir ranges , 4 ) the vast leeway for model tailoring , offered by the disordered structure of coals and kerogens , as opposed to the very restrictive aromaticity of pahs , 5 ) the credibility of the model material , because it is so easily found in nature and essentially made of only the four most abundant elements in space ( after he ) , and 6 ) the fact that this material has already been so thoroughly studied in the laboratory that even its diversity is mostly understood . the issue of the excitation mechanism is intimately linked to the issue of modelling the observed spectra . by invoking chemiluminescence on large particles , one avoids high dust temperatures and the attendent puzzles of varying band wavelengths and widths which are not observed in the sky . -the need for a library of complete celestial ir spectra ( @xmath112 to @xmath1116 @xmath0 m ) , each associated with a known environment , differing from the others by the intensity of the radiation field or the nature of the object : rn , ppn , pn , pdr , gc , etc . ; this would help understand , by analogy with coals and kerogens , the evolution of is dust from inception in the envelopes of agb stars to its consumption in protostars ; the coal / kerogen model can set a frame for the discussion of these and other issues . however , keeping in mind that these terrestrial materials were forged under special physical and chemical conditions , the model will no doubt have to be amended according to observations , as the debate goes on . i am indebted to j. conard , j .- n . rouzaud ( cnrs / orleans ) , and m. cauchetier , o. guillois , i. nenner , c. reynaud ( cea / drecam , saclay ) for a fruitful collaboration , and to many others for enlightening discussions .	the recent influx of high - quality infrared spectroscopic data has prompted an extensive reassessment of various laboratory models in comparison with the observed uibs ( unidentified infrared bands ) . as a result , significant modifications were brought to the original paradigms . the focus here is on the evolution of the coal model , characterized by 1 ) a shift towards less aromatic materials of the same family ( kerogens ) , 2 ) the introduction of a new excitation mechanism ( chemiluminescence ) , based on the capture of hydrogen atoms by carbonaceous dust . both developments are intended to accomodate observations from a larger range of dust environments and evolutionary stages . this leads to a more quantitative description of dust composition and structure , and a better understanding of its history . in short , according to the present model , the evolution of dust from inception in the circumstellar shells of agb stars , through strong interstellar radiation fields , to consumption in protostars , is approximately mimicked by progressively more aromatic materials , starting from the young and mostly aliphatic kerogens , through more and more mature coals , to the final stage of polycrystalline graphite . a similar family of materials is obtained in the laboratory by annealing in vacuum up to or beyond 3000 k. the composition , structure and ir spectrum of these materials are extensively documented in the geology literature . the fundamental characters of this model , viz . chemical and structural disorder , and diversity of chemical bondings , naturally point to ways of further tailoring , in order to fit particular observations more closely .
the puzzle of glass - forming systems has remained sufficiently elusive over the years such that even the puzzle pieces themselves have changed shape . for example , the puzzle piece of the lack of a growing lengthscale may now have to be modified since a growing lengthscale can perhaps be extracted from a higher - order correlation function @xcite . a less recent change in puzzle pieces is the distinction by angell between fragile and strong glasses , where certain effects are more dramatic in fragile glasses than in strong ones @xcite . one piece of the puzzle that has remained constant over the years , however , is the dramatic slowing down of the dynamics of the particles near the glass transition . more precisely , a supercooled liquid s viscosity can increase by fourteen orders of magnitude as the temperature is decreased near a `` working '' definition of the glass temperature @xcite . two main phenomenological models for this dynamical slowing down have emerged over the years : mode - coupling theory and kinetically constrained models . we will not focus on mode - coupling theory here and simply refer the reader to several excellent reviews @xcite . as for the second phenomenological approach , kinetically constrained models , the goal is to understand whether glassy dynamics can be understood as arising from steric constraints on the particles alone @xcite . one of the simplest such examples is the kob - andersen model @xcite . it is motivated by the caging of particles , ultimately observed in larger scale systems such as colloidal glasses @xcite . the kob - andersen model is a hard - core lattice gas model , but with the constraint that a particle can only hop to an adjacent site if and only if it has less than a certain number of neighbors , @xmath0 , both before and after the move . early simulations of the kob - andersen model on the hypercubic lattice for relevant values of @xmath0 appeared to find a dynamical phase transition at a nontrivial critical density @xcite . however , subsequent mathematically rigorous results found that the phase transition does not occur until the fully packed state @xcite . this corresponds to a zero - temperature glass transition . tbf proved this by showing that at any monomer density , there were mobile cores that could diffuse at sufficiently long time scales . while the hypercubic version of the kob - andersen model does not exhibit a finite - temperature glass transition , the mean field version does @xcite . since it is still up for debate whether or not mean field is relevant for physical systems , one can ask whether or not there exists another finite - dimensional , kinetically constrained model that exhibits a finite temperature glass transition . while the kob - andersen ( and the fredrickson - andersen @xcite ) models , are elegant in their simplicity , there are indeed two more involved kinetically constrained models in two dimensions that can be proven to exhibit a finite - temperature glass transition @xcite . these two models have been dubbed the spiral model @xcite and the sandwich model @xcite . both models exhibit an unusual phase transition in that the fraction of frozen particles jumps discontinuously at the transition , typical of a first - order phase transition . however , as the transition is approached from below , there exists a crossover lengthscale that diverges faster than a power law in @xmath1 . the crossover length , @xmath2 , distinguishes between squares of size @xmath3 , which are likely to contain a frozen cluster , and squares of size @xmath4 , for which the probability of containing a frozen cluster is exponentially unlikely . given this combination of a discontinuity in the fraction of frozen particles , and a faster than power law diverging length scale , the transition has unique characteristics @xcite . models such as the spiral and sandwich models are proof in principle that further exploration of kinetically constrained models in finite - dimensions may be fruitful , in particular , because should an ideal glass transition exist , it may indeed be of unusual character . more specifically , the edwards - anderson order parameter should be discontinuous at the transition , yet accompanied by rapidly diverging time scales @xcite . kinetically constrained models can be related to models of correlated percolation , which are models in which each site is initially occupied with an independent probability @xmath5 , as in normal percolation , but then correlations are induced through some culling rule that is , sites can only be occupied if certain conditions on their neighboring sites are met . the two types of models are related , with the immobile particles of the kinetically constrained models corresponding to the stable occupied particles of the correlated percolation models . the simplest model of correlated percolation is @xmath6-core percolation , in which an occupied site must have at least @xmath6 occupied neighbors @xcite . occupied sites that do not satisfy this stability requirement are removed , and this condition is applied repeatedly , until all remaining sites are stable . this model can be mapped to the fredrickson - andersen model @xcite . one might think that kinetically constrained models and correlated percolation systems would be easy to numerically simulate , and that construction of rigorous proofs would be nothing more than an interesting problem for mathematicians . however , while simulating these systems is easy , extracting their properties in the infinite system size limit is much harder . just as with the kob - andersen model , initial numerical simulations of @xmath6-core percolation for certain values of @xmath6 found evidence of first - order phase transitions at nontrivial critical densities , and of second - order phase transitions in a different universality class than normal percolation @xcite . however , subsequent mathematically rigorous analyses found that the critical point is @xmath7 for those @xmath6 @xcite . because the critical point only approaches unity very slowly in the limit of infinite system size , the simulations on finite - size systems were misled as to the location of the true critical point , thereby highlighting the importance of rigorous results for these models as well . for a review of @xmath6-core percolation , see ref . @xcite . there is another finite - dimensional system exhibiting an unusual transition , at least numerically . it is the jamming transition in repulsive soft spheres . numerical simulations of soft spheres show a critical point at which the average coordination number jumps discontinuously to a universal , isostatic value . but quantities such as the shear modulus and the deviation of the average coordination number from its isostatic value show a nontrivial power law behavior in the vicinity of the critical point @xcite . recent experiments on two - dimensional photoelastic beads support this notion of a mixed transition @xcite . interestingly , it has been conjectured that the physics of granular systems , colloidal systems , and glassy systems are of a similar character @xcite . the mean field results of not one but several correlation percolation models corresponding to glassy physics support this notion quantitatively . furthermore , experimental evidence of caging in another two - dimensional granular system also supports this notion @xcite . the question of finite - dimensional glassy models being quantitatively similar to the repulsive soft sphere system is still being investigated . certainly the spiral and sandwich models show that , qualitatively , one can have a glassy system exhibiting an unusual finite - dimensional transition . however , they do not appear to be in the same class as the jamming system , since the order parameter exponent is unity just above the transition in the jamming percolation models , but is one - half in the jamming system . to explore the possible link between jamming and glassy systems in terms of a finite - temperature glass transition , tbf initially introduced the knights model , a model of correlated percolation similar to the spiral model @xcite , and called it a model of jamming percolation . in fact , the spiral , knights , and sandwich models are all models of jamming percolation . in this paper , we expand on our earlier work ( ref . @xcite ) , in which we introduced the sandwich model , by presenting the details of the proof of an unusual transition in this model . this proof is based on modifying the proof developed by tbf in refs . @xcite for the spiral model ( although originally misapplied to the knights model ) . we also introduce further generalizations of models of jamming percolation to demonstrate that the phenomenon of a finite - dimensional transition is indeed somewhat generic . in doing so , we show that the tbf proof only gives a rigorous derivation of these novel properties if a `` no parallel crossing '' rule holds . this rule says that two similarly - oriented directed percolation chains can not cross without having sites in common . the effect of this rule is that one directed percolation - like process can not be used to locally stabilize the other . for models such as the sandwich model , which satisfy this `` no parallel crossing '' rule , but fail a `` no perpendicular crossing '' rule , the tbf proof works only with some modification . the methods described in this paper can be used to understand for which correlated percolation models a proof along the lines of the tbf proof can be used to show a glassy transition , and for which models they can not . given that there are two very detailed papers on the tbf proof @xcite , we will refer to them quite often , as opposed to making this paper self - contained . finally , we will discuss connections between jamming percolation and force - balance percolation , another correlated percolation model inspired by granular systems , where numerical evidence points toward an unusual transition in finite dimensions @xcite . we consider a class of models that generalizes the knights model , the earliest of the jamming percolation models . the class of models is defined on the two - dimensional square lattice . initially , each site is occupied with an independent probability @xmath5 . each site has four neighboring sets , and each set contains two sites . the four sets are labelled as the northeast , northwest , southeast , and southwest neighboring sets . ( the sites in those sets only lie in precisely those compass directions for the knights model , but we continue to label the four sets in this manner for all our correlated percolation models until section [ sec : the implications of property b ] . ) to be stable , an occupied site must either have ( 1 ) at least one northeast neighbor and at least one southwest neighbor , or ( 2 ) at least one northwest neighbor and at least one southeast neighbor . all other occupied sites are unstable , and are vacated . this culling process is then repeatedly applied sites that were previously stable may become unstable by earlier cullings until all remaining sites are stable . the neighboring sets of the original knights model introduced by tbf @xcite are shown in fig . [ fig : knights ] . fig . [ fig : sandwichmodel ] shows the neighboring sets in the sandwich model , which we introduced in @xcite , and fig . [ fig : spiralmodel ] shows the neighboring sets in the spiral model , which tbf introduced in @xcite . in these models , stable sites must either be part of a chain running from the northeast to the southwest , or part of a chain running from the northwest to the southeast . in the final configuration , all sites must be stable , so any chain must either continue forever ( to the boundary ) , or terminate in a chain of the other type ( this latter case is called a `` t - junction '' ) . any sites left after the culling procedure must thus be connected by an infinite series of chains ( or , in a finite system , connected by a series of chains to the distant boundary ) , so asking for the critical probability at which an infinite cluster first appears is the same as asking for the minimum probability at which some sites remain unculled in the infinite size limit . if sites could only be stable by having northeast and southwest neighbors , then every stable site would be part of a northeast - southwest chain on a particular sublattice the sublattice extending to the northeast of the site for the sandwich model is shown in figure [ fig : sublattice.sandwich ] . chains on this sublattice are isomorphic to infinite chains in directed percolation , which has a well - studied second - order phase transition @xcite . however , in these models there is an additional mechanism by which sites can be made stable that is , by having northwest and southeast neighbors . adding an extra way for a sites to be stable can only possibly depress the critical probability , so we immediately see that for these models @xmath8 . it will turn out to be useful to divide these models into classes based on two properties . we define a model as having `` no parallel crossing '' property if whenever two northeast - southwest chains ( or two northwest - southeast chains ) intersect , they must have sites in common we abbreviate this a `` property a. '' and we define a model as having a `` no perpendicular crossing '' property , or property b , if whenever a northeast - southwest and northwest - southeast chain intersect , they must have sites in common . table [ table : properties ] shows which properties each of the three models possesses : @xmath9 examples of where the properties fail for each model are shown in figure [ fig : crossings ] . our analysis here is based to a large extent on the claimed tbf proof of a glassy transition for the knights model . their proof consisted of two parts . first , they claimed to show that the critical point of the knights model is exactly the same as that for directed percolation . second , once this was done , they were able to use well - known results on directed percolation ( assuming a well - tested conjecture about anisotropic rescaling in directed percolation ) to show that this model has a glassy transition specifically , they were able to find structures with a finite density at the critical point of directed percolation , and to show that just below this critical point , the crossover length and culling times diverged . a short version of their proof appeared in refs . @xcite , with more detailed explanations in ref . @xcite and @xcite . our analysis of this more general class of models shows that the tbf proof that @xmath10 is only valid for models satisfying property a ; so it works for the sandwich and spiral models , but fails for the knights model . the second part of their proof , showing a glassy transition , implicitly assumes property b. the spiral model exhibits property b and hence the tbf method of proof carries through @xcite . however , we show that the proof can be modified to work for models that fail to have property b. the spiral _ and _ the sandwich models thus have provably glassy transitions , while the knights model does not . we sketch the tbf proof that @xmath11 , which will let us understand why property a is sufficient , and most likely necessary to the result . the key to the tbf proof is to show that voids ( clusters of empty sites ) of particular shapes have a finite probability of growing forever . for example , for the diamond - shaped void in the sandwich model , shown in figure [ fig : sandwich.void ] , if the key site labelled @xmath12 is vacant , it will trigger the removal of all the sites marked with stars , increasing the void size by one . ( the corresponding void for the knights model appears as figure 1c of @xcite . ) if this process repeats forever , with such key sites repeatedly removed , this void will grow to infinity . the tbf proof for the knights model is based on the claim that the key sites of the knights model , located at the corners of octagonal voids of size @xmath13 , can only be stable if part of a directed percolation chain of @xmath14 . once this claim is granted , the rest of the proof is straightforward . for @xmath15 , such long chains are exponentially suppressed , and thus for a large void , the vertices at the corners of the void are exponentially likely to be culled . summing up the relevant probabilities , this results in a finite probability that the void will grow to infinity . for an infinite lattice , it is thus certain that there will be at least one void that grows to infinity , showing that all sites are culled below @xmath16 . since @xmath17 , this is supposed to show that @xmath11 . the claim that sites at the corners of voids can only be stable if part of a long chain of @xmath14 is only valid for models with property a. a counterexample to this claim for the knights model can be seen in fig . [ fig : counterexample ] . this counterexample also shows why property a is necessary and sufficient for this claim to be valid . to be stable , the site must be part of a northeast - directed chain ; pick the lowest northeast - directed chain coming out of the corner site @xmath12 . if that chain stops before reaching length @xmath14 , it must terminate in a northwest - southeast chain . since the northeast chain is not as long as a wall of the void , the new southeast - directed chain will eventually hit the void ( thus resulting in the culling of all chains , and the corner site @xmath12 ) , unless it hits a t - junction . that new t - junction results a second northeast - southwest chain , which will eventually reach the first northeast chain , as in figure [ fig : counterexample ] . for models with property a , the two chains will intersect , contradicting the original assumption that we chose the lowest northeast - directed chain out of @xmath12 . thus , by contradiction , for models with property a , the first northeast chain must be @xmath14 for @xmath12 to be stable , and it indeed follows that @xmath10 . what about for models such as the knights model , that lack property a ? is it possible that despite this counterexample to the claim , @xmath18 is actually equal to @xmath19 , for some other reason ? while we do not have a mathematically rigorous proof that @xmath20 , we present an argument that the two are almost certainly unequal . we present our arguments in the context of the knights model , but they generalize to other models that lack property a. consider the substructure in fig . [ fig : sub1to2 ] . all sites in it are stable under the knights model culling rules , except for the two sites at its ends , and those sites will become stable if the substructure is attached between two northeast - southwest chains . furthermore , there is no northeast - southwest chain internal to the substucture connecting the two ends . the substructure is internally stabilized by northwest - southeast links . this means that the substructure can act as a `` rest stop . '' northeast - southwest chains can have breaks in their paths , connected by this substructure . just below @xmath16 , long northeast - southwest chains almost form an infinite structure . they are almost linked , so a few extra connections , through these substructures , should create an infinite cluster even below @xmath16 . so we expect that @xmath21 . we can make the argument more formal by considering the following modification of the directed percolation problem , which we call jumping directed percolation . as with normal directed percolation , we occupy sites on the square lattice with probability @xmath5 , and connect each site with directed bonds to its neighbors to the north and east . however , now we define an additional way for sites to be connected . we divide the lattice into blocks of size 9x9 , and for each block , if the two hollow squares of sites shown in fig . [ fig : jumpperc ] are occupied , we with probability @xmath22 connect the two hollow squares with a directed bond from the southwest square to the northeast square . the critical point of this model is a function of @xmath22 : @xmath23 . by repeating fig . [ fig : sub1to2 ] three times , to create diamonds connecting sublattice # 1 to # 2 to # 3 and then back to # 1 , we obtain a structure that links two separated diamonds on sublattice # 1 , through sites in the other two sublattices . so if we restrict ourselves to looking at sites in sublattice # 1 , and the directed percolation structures on that sublattice , sites that appear disconnected may be connected by these sites in sublattices # 2 and # 3 . the structure obtained by repeating fig . [ fig : sub1to2 ] three times has 24 sites in sublattices # 2 and # 3 , each of which is occupied with probability @xmath5 , and has 24 sites in sublattice # 1 . the sites in sublattice # 1 in this repeated structure map onto the occupied sites in fig . [ fig : jumpperc ] . so with @xmath24 , @xmath22 gives the probability of having appropriate `` hidden '' occupied sites in sublattices # 2 and # 3 that connect and make stable the two hollow squares . infinite chains in the jumping directed percolation model are infinite stable clusters in the knights model , and thus @xmath25 . however , since jumping directed percolation is just directed percolation with an extra connection process , it is reasonable to expect that @xmath26 for all @xmath27 , implying @xmath21 . while this argument is not mathematically rigorous , it is strongly suggestive , particularly when we recall previous results on enhancements in percolation by aizenman and grimmett @xcite . their work showed that if percolation on a lattice was `` enhanced '' by adding , for specified subconfigurations , extra connections or occupations with probability @xmath22 , this would _ strictly _ decrease the critical probability , for any @xmath27 , so long as the enhancement was essential . essential enhancements were defined as those such that a single enhancement could create a doubly - infinite path where none existed before . see ref . @xcite for a more rigorous and precise statement of the results on enhancements , and ref . @xcite for a general discussion of enhancements . the results of ref . @xcite were obtained for undirected percolation , so are not directly relevant for the jumping directed percolation model considered here , but they are analogous enough to strongly suggest that @xmath26 for all @xmath27 . it is difficult to see how such adding such a new route for paths to infinity could leave the critical probability completely unchanged . for models satisfying property a , we have @xmath10 , but we still need to check that the tbf proof that the transition is glassy ( discontinuous with a diverging crossover length ) is valid . the tbf proof of a glassy transition implicitly assumes that the knights model has property b. for example , to show discontinuity , they construct a configuration that has a finite density at @xmath16see fig . 2b of @xcite . this figure , and others like it , implicitly assume property b , because they are based on drawing overlapping rectangles in independent directions , and assuming that if paths in these intersecting rectangles cross , they must stabilize each other ( form a t - junction ) . the resulting frozen structure is shown on the left side of figure [ fig : twofrozenstructures ] . however , if a model lacks property b , the paths can cross without stabilizing each other . the knights model does not satisfy property a , so whether or not it satisfies property b is a moot point . but what about the sandwich model , which satisfies property a , but not property b ? the tbf proof as it stands is not immediately valid in these cases . nevertheless , it turns out that for such models , the tbf proof can be made to work by a modification of their structures . the basic idea of the modification is as follows . the tbf proof of a glassy transition is based on drawing structures consisting of sets of overlapping rectangles , showing that there is a sufficiently high probability ( finite for the proof of discontinuity , and approaching 1 for the proof of diverging crossover length ) that each rectangle has a spanning path in the desired direction , and then using property b to conclude that the intersecting paths form a frozen cluster . for models that lack property b , we use the same figures as in the tbf proof ( e.g. figs 2a and 2b of ref . @xcite ) , but pick the rectangle sizes large enough that there is a high probability that each rectangle has _ multiple _ spanning paths ( @xmath28 for a rectangle of @xmath14 ) , each occurring in a disjoint parallel subrectangle . then in each place where the tbf proof assumes a t - junction based on property b , we will have a northeast ( northwest ) path crossing many northwest ( northeast ) paths . the probability that no t - junction occurs turns out to decay exponentially with the number of northwest ( northeast ) paths . in figure [ fig : twofrozenstructures ] we show how the discontinuity structure of the tbf proof ( from fig . 2a of ref . @xcite ) is modified by this procedure . to implement these ideas , we need to modify proposition 5.1 of ref . @xcite , which says that sufficiently large rectangles of size @xmath29 are very likely at the critical point to have chains connecting the sides of length @xmath30 : * proposition 5.1 * ( from ref . @xcite . ) : + for @xmath31 there exists @xmath32 s.t . + @xmath33 here @xmath34 is the the dynamical exponent , which has been numerically found to be approximately 0.63 in two dimensions @xcite . the tbf proof of this proposition assumes a conjecture that there is a well - defined @xmath35 ( conjecture 3.1 of ref . we replace their proposition 5.1 with the following proposition , which instead states that we are likely to have @xmath28 connecting chains in disjoint parallel subrectangles : * proposition 100 * : + for @xmath31 there exists @xmath32 and @xmath36 s.t . + @xmath37 we divide the box of size @xmath13 by @xmath30 into @xmath38 parallel disjoint subrectangles , each of size @xmath39 . assuming the conjecture of anisotropic scaling in directed percolation ( conjecture 3.1 of ref . @xcite ) , each subrectangle has a probability @xmath40 of having a path connecting the two sizes of length @xmath41 , contained within that subrectangle . the expected number of crossings is @xmath42 , and for any @xmath43 the probability of having less than @xmath44 crossings decays exponentially in @xmath45 . the tbf proof of discontinuity shows that for certain structures of rectangles , there is a nonzero probability that each rectangle has a suitable `` event , '' and that , assuming property b , the existence of each event ( a rectangle crossing ) results in a stable structure at the critical point . now , with our modified proposition , each `` event '' is the presence of multiple crossings ( in disjoint parallel subrectangles ) in each rectangle , rather than single crossings . without property b , this does not guarantee a frozen cluster . however , we see in appendix [ sec : independence ] that when a northeast ( northwest ) path crosses @xmath46 northwest ( northeast ) paths , in @xmath46 disjoint subrectangles , the probability of not forming a single t - junction decays exponentially in @xmath46 . this is physically obvious , since for large subrectangles , the probability of each t - junction in each subrectangle is essentially independent ; however , since the probabilities are not truly independent , more work is needed to make this rigorous . the details of the proof are relegated to appendix [ sec : independence ] . more generally , the arguments in appendix [ sec : independence ] show that we can treat the probabilities of t - junctions in different subrectangles as independent , when establishing an upper bound on the probability . we will use this throughout this section to multiply such probabilities as if they were independent . assuming property b , the tbf proof shows that the probability of having suitable events in each rectangle is nonzero at the critical point . we now need to show that , even without property b , this results in a finite probability of an appropriate set of t - junctions . in the tbf discontinuity structure , shown in fig . 6 of ref . @xcite , the rectangles are labelled @xmath47 , @xmath48 . each rectangle @xmath47 is twice as large as the rectangles @xmath49 . so the `` suitable events '' of proposition 100 give @xmath47 at least @xmath50 crossings parallel to its long direction , where @xmath6 is some positive constant . the arguments in appendix [ sec : independence ] show that this results in a t - junction with probability @xmath51 , for some positive @xmath52 and @xmath6 . starting at the origin , the probability of forming appropriate t - junctions off to infinity can then be seen to be @xmath53 for some positive @xmath54 , @xmath55 , and @xmath6 . this product converges to a positive number , so the transition is proven to be discontinuous ( subject to assumption of the well - tested conjecture of an anisotropic critical exponent in directed percolation ) . the proof of the diverging crossover length can be made to avoid the assumption of property b by a similar modification of the tbf structures . again , we begin by repeating the tbf structures , with the set of parallel rectangles in figure 2a of ref . @xcite . in that picture , if every rectangle has a spanning path , _ and _ the paths all intersect , there will be a spanning frozen cluster . tbf consider the case where each rectangle has sides of order the directed percolation parallel correlation length , @xmath56 . they then show that @xmath57 and @xmath58 can be chosen such that if the system size is @xmath59 $ ] , the probability that each rectangle is occupied by a spanning cluster approaches @xmath60 as @xmath31 , @xmath61 . if property b were to hold , this would result in t - junctions that would create a frozen structure , and show that the crossover length diverges as @xmath61 . we no longer have property b ; but instead , by replacing proposition 5.1 with proposition 100 , we can choose the rectangle sizes such that each rectangle is occupied by `` many '' spanning clusters ( with `` many '' defined by proposition 100 ) . given this , we can start at an arbitrary rectangle , and then work our way out , looking for t - junctions to create a spanning frozen structure . we will only fail to create a frozen structure if at some point we reach intersecting rectangles where one spanning path in one rectangle crosses many spanning paths in the other rectangle , but without creating a t - junction . by the arguments in appendix a , the probability of this occurring decays exponentially in @xmath62 for some @xmath32 . so even with @xmath63 intersections , the probability that we ever fail in this process goes to 0 as @xmath31 and @xmath61 , and we are essentially guaranteed a frozen structure . this shows that the crossover length diverges as we approach the critical point , with the same lower bound that tbf found . the models we have discussed so far have two possible ways in which a site can be stable , and varying neighboring relations for the culling relations . however , we can also define generalizations in which there are three or more ways in which a site can be stabilized . for example , for the pinwheel model , shown in figure [ fig : pinwheel ] , the condition for a site to be stable is that it ( 1 ) have neighbors in both the sets a and b , or ( 2 ) have neighbors in both the sets c and d , or ( 3 ) have neighbors in both the sets e and f. this gives a site three possible directions for stabilizing chains . similarly , figure [ fig : pinwheelplus ] shows a model in which there are four possible directions for a site to be stable , such that there is an extra `` or '' : for the sets g and h. we denote this model the 8-spiral model . despite the extra ways in which sites can be made stable , the tbf proof of a glassy transition is still valid , because property a holds for both the pinwheel model and the 8-spiral model . that is , in both of these models if two a - b chains ( or two c - d chains , or two e - f chains , or two g - h chains ) cross , they must have sites in common . this turns out to be sufficient to show that there will be a void that grows forever in the infinite system limit . for the sandwich and spiral models , we needed to show that a stable site at the corner of a diamond - shaped void has to be part of a directed percolation - like chain ( dp - like chain ) of order the size of the void . for the pinwheel model , we consider hexagonal voids , and show that stable sites at the corners of these voids must be `` associated '' with a long dp - like chain , where `` associated '' will be defined by the construction below . then , just as for the sandwich and spiral models , for @xmath64 , long dp - like chains are exponentially suppressed , giving voids a finite probability to grow forever , showing that the infinite system is empty for @xmath64 . consider the site @xmath12 at the corner of the hexagonal void in figure [ fig : pinwheel.void ] . look at all a - b chains coming out of @xmath12 , and pick the lowest possible chain ; in other words , look for successive a neighbors , and if a site has two possible a neighbors , pick the lower one . if that chain reaches the dashed line in figure [ fig : pinwheel.void ] , we have a dp - like chain of order the size of the void , and are done . otherwise , this a - b chain must terminate either in a c - d chain , or an e - f chain . if it terminates in an e - f chain , that chain must terminate in a c - d chain , which must then cross the original a - b chain . ( note that the e - f chain can not terminate in an a - b chain , since by the `` no parallel crossing '' rule , the new a - b chain would intersect the first a - b chain , and contradict our assumption that we chose the lowest a - b chain coming out of the site @xmath12 . ) so if the a - b chain coming out of @xmath12 does not reach the dashed line , it must either terminate in a c - d chain or cross a c - d chain(the latter case is shown in figure [ fig : pinwheel.void ] ) . pick the lowest of all the c - d chains that cross or intersect our a - b chain . this c - d chain must reach the dashed line , using the same logic as before ( if it terminated in an a - b chain , that would intersect the first a - b chain , and contradict the assumption that we chose the lowest a - b chain coming out of @xmath12 ; while if it terminated in an e - f chain , that e - f chain would have to turn into either an a - b or c - d chain before reaching the void , again resulting in a contradiction . ) since the a - b and c - d chains that we have constructed cross , and together include both @xmath12 and the dashed line , at least one of the chains must be of order the size of the void . a similar construction can also be used for the 8-spiral model using a diamond - shaped void . having established where the critical point is , we follow tbf , and as before consider an infinite sequence of two types of rectangular regions , @xmath65 ( see fig . [ fig : twofrozenstructures ] ) , one of which contains ab paths , and the other of which contains ef paths . ( any pair of types of paths ab / cd , ab / ef , or cd / ef are permissible , as are any of the 6 possible pairs for the 8-spiral model . ) again , the sequence is constructed such that the ab paths and the ef paths are mutually intersecting with a frozen backbone that contains the origin , and the tbf proof simply carries through with the additional modification we have introduced for models that do not obey property b. all in all , since having more neighboring relations gives more ways for an occupied site to be stable ( without depressing the dp critical point ) , the tbf constructions of discontinuous percolation structures at the critical point simply carry through . just as for the sandwich model , the tbf proof needs to be modified to deal with the lack of property b. one needs at least two intersecting rectangular regions in which the probability for two `` transverse '' blocking directions each undergoing a directed percolation transition independently is nonzero . consideration of the `` no parallel crossing '' rule shows that for higher - dimensional generalizations of the knights model , the tbf proof can not be generalized in a straightforward manner to show provably glassy transitions . we will show that the``no parallel crossing '' rule never holds , so the critical point is always depressed below that of directed percolation . to be specific , suppose that in three dimensions we have disjoint neighboring sets @xmath66 , @xmath67 , @xmath68 , and @xmath69 , generalizing the northeast , southwest , northwest , and southeast sets of the knights model ( see figure [ fig : knights ] ) . each set should consist of three linearly independent vectors , and the sets @xmath66 and @xmath67 should be opposite of each other , as should the sets @xmath68 and @xmath69 . then , just as in the knights model , the culling rule is that a stable site should have occupied neighbors in both the sets @xmath66 and @xmath67 , or in both the sets @xmath68 and @xmath69 . then , just as for the sandwich and spiral models , there are two directed percolation processes by which a site can be made stable ( @xmath70 chains and @xmath71 chains ) , and one might think that for an appropriate set of neighboring relations the tbf proof could be used to show a glassy transition . however , it turns out that because these models never satisfy three - dimensional generalizations of the `` no parallel crossing '' rule , the critical point is always depressed below that of three - dimensional directed percolation , and the tbf proof can not be directly generalized for these models . property a says that two chains running in similar directions can not cross without having sites in common . for certain two - dimensional models , such as the sandwich and spiral models , these property can be required by the topology and neighboring relations . however , in three dimensions , the topology always makes it easy for two directed chains to miss each other , and so no three - dimensional generalization of property a can be satisfied , regardless of the neighboring relations . furthermore , if the two chains miss each other , then the buttressing of each type of chain does not occur and the resulting transition may be continuous . this rough argument can be formalized by showing that three - dimensional generalizations of the knights model always have substructures such as the one shown in figure [ fig : sub1to2 ] . that is , it is always possible to find substructures that have no long @xmath70 chains connecting their ends , but which can join up two @xmath70 chains and stabilize their ends . then , by the arguments in section [ sec : critical.point ] , we should expect these to depress the critical point below that of three - dimensional directed percolation . specializing to three dimensions for convenience , let @xmath72 consist of three linearly independent 3-vectors , and @xmath73 consist of three linearly independent 3-vectors with @xmath74 . also , define @xmath75 and @xmath76 . for an occupied site @xmath77 to be stable , it must have either ( 1 ) occupied neighbors from both @xmath78 and @xmath79 or ( 2 ) occupied neighbors from both @xmath80 and @xmath81 . the first condition we denote the a - b condition , the second , the c - d condition . if we only enforced the a - b condition , then we would just have three dimensional directed percolation ( modulo finite clusters ) . however , for the model defined above , there is an extra way to be stable , resulting in @xmath82 . consider a _ finite _ structure with the following properties : ( 1 ) all occupied sites are stable under the culling rules except occupied sites @xmath83 and @xmath84 , ( 2 ) @xmath83 has an occupied neighbor in @xmath79 and @xmath84 has an occupied neighbor in @xmath78 , and ( 3 ) there is no ab path connecting @xmath83 and @xmath84 , but the structure is stable because there exists a path where at least one occupied site is stable under the c - d condition . we relegate to appendix b the proof of the existence of such a finite structure . with the finite structures defined above , some occupied sites that were unstable under the a - b condition now become stable . slightly below @xmath85 , the system is about to percolate using the a - b condition alone . the substructures act as extra , local bonds , joining up long a - b paths and pushing the system above the critical point , just as in the knight model . again , arguments similar to these have been made rigorous by aizenmann and lebowitz in the case of undirected percolation , and perhaps can be made rigorous in the case of directed ( oriented ) percolation . the discovery of a two - dimensional percolation transition , where the sudden emergence of a discontinuous backbone coincides with a crossover length diverging faster than a power law , is recent , and of great interest for glassy systems , jamming systems , and phase transitions in general . unusual transitions have been found previously in mean field systems of a slightly different nature , but not in finite dimensions . and while the finite - dimensional transition is discontinuous , it is not driven by nucleation , as with ordinary discontinuous transitions , but instead by a scaffolding of many tenuous directed percolation paths occurring simultaneously to form a bulky structure . we have shown that property a is required for the proof that @xmath86 , but property b is not . all that is needed to prove that the transition is discontinuous ( once @xmath86 is established ) is a finite probability for two transverse percolating structures to intersect , to prevent each other from being culled . therefore , one can construct other models , such as the pinwheel and 8-spiral models , that exhibit a similar transition in two dimensions . the phenomena is not as specific as might seem at first glance . however , such a buttressing mechanism in dimensions higher than two is more difficult because it is more difficult for percolating _ paths _ to intersect and form a buttressing , bulky structure . models like the knights model , where @xmath87 is most likely less than @xmath85 , provide physicists , mathematicians , and computer scientists with a motivation to study new models of correlated percolation models that are not isomorphic to directed percolation , but quite possibly in the same universality class . once this avenue is pursued further , one can then easily extend the class of models for which a finite - temperature transition can be rigorously shown . to begin , it would be interesting to consider a directed percolation model in two dimensions where the number of nearest neighbors is greater than two . for example , if the number of nearest neighbors was increased to four , would the percolation transition still be in the same universality class as directed percolation ? if so , as is presumably the case , then one could construct a jamming percolation model with sets larger than two sites . these jamming percolation models would then be isomorphic to the next - neighbor directed percolation models and then one could use results from directed percolation to prove a percolation transition . there exists another class of correlated percolation models called force - balance percolation . the first model in this class was defined in ref . other force - balance percolation models are currently being constructed and studied @xcite . the force - balance percolation models differ from the jamming percolation models in that ( 1 ) the sets , such as a , b , etc . , are overlapping and ( 2 ) the `` or '' between pairs of sets is changed to `` and '' . given these differences , the methods of proof used here can not be easily applied . numerical results indicate that the transition is discontinuous with a nontrivial `` correlation length '' exponent , indicating that the transition may not the garden - variety discontinuous transition . the force - balance models are perhaps less artificial in that they mimic force - balance by requiring that an occupied site ( i.e. a particle ) have occupied neighbors to its left and right , as well as its top and bottom , in order to be stable . however , little has been rigorously proven about them . to make progress along these lines would be useful . what about the lack of finite stable clusters in models of jamming percolation ? recent numerical work on another correlated percolation model , @xmath6-core percolation , with @xmath88 on the four - dimensional hypercubic lattice , appears to exhibit an _ , discontinuous percolation transition driven by nucleation @xcite . finite clusters exist in this model , unlike in the jamming percolation models . therefore , in the jamming percolation models , there can be no surface tension between the percolating and nonpercolating phases , which is typical of an ordinary discontinuous transition . if finite clusters are allowed in a correlated percolation model , one might guess that the unusual nature of the transition would be destroyed . however , finite clusters , other than individually floating particles , do not appear in the jamming transition of granular particles . otherwise , the packing would not be static . so it is unclear whether the existence of finite clusters pertains to the jamming of granular particles . this is also the case for the glass transition . if more of an analogy between jamming and models of jamming percolation is to be made in finite dimensions ( setting aside the matter of the critical dimension of jamming ) , a model where the fraction of sites participating in the infinite cluster increases smaller than linearly just above the transition must be found . furthermore , the existence of a jamming percolation model with a universal jump in the number of occupied sites at the transition that `` naturally '' emerges as opposed to being externally imposed @xcite , is yet another necessary quest if a jamming percolation model of jamming is to be found . finally , models of correlated percolation , such as the sandwich model , tell us that there do indeed exist kinetically constrained models of glassy dynamics that exhibit unusual phase transitions in finite dimensions . therefore , this avenue of exploration for understanding possible finite - temperature glass transitions in finite - dimensions remains open . since our work helps to clarify which jamming / correlated percolation models can be rigorously shown to have an unusual finite - temperature glass transition with a particular set of properties , other models exhibiting possibly other unusual behaviours can hopefully be more easily developed in the near future . in this appendix we justify the claim made in section [ sec : the implications of property b ] that given @xmath46 crossings , the probability that no crossing results in a t - junction decays at least exponentially in @xmath46 . this would be immediately true if each crossing resulted in an independent probability of a t - junction . so what we show is that these crossings , by occurring in disjoint subrectangles , can be effectively treated as independent ( in establishing an upper bound on the probability ) . the relevant picture is shown for @xmath89 in fig . [ fig : multiplecrossings ] . there are @xmath46 disjoint rectangles , labeled by @xmath90 , @xmath91 , each of which has at least one northwest spanning path . we call this event @xmath92 , and conditionalize upon the occurrence of @xmath92 . each northwest path must cross the northeast spanning path , but without property b , these crossings do not necessarily result in t - junctions , where by a t - junction we mean specifically a site in common between the paths that stabilizes the northeast part of the northeast path . let @xmath93 be the event that at least one crossing in rectangle @xmath90 forms a t - junction . now for a configuration with @xmath94 , look at the sites in the vicinity of the crossing ( if there is more than one crossing , we choose one by an arbitrary ordering of possible crossing locations ) . restricting ourselves first to the sandwich model , fig . [ fig : missedcrossing ] shows the only way that @xmath93 can fail to happen . the site labeled by a star must be vacant , and if that site is made occupied , the new configuration is in @xmath93 . so a local change in the vicinity of the crossing can always create a t - junction . this local change induces a mapping @xmath95 from the set of states with @xmath94 to a subset of the set of states with @xmath93 . the mapping is one - to - one onto this subset , and for any state @xmath96 with @xmath94 , the probability of the configuration @xmath97 is @xmath98 times the probability of the configuration @xmath96 , where @xmath5 is the site occupation probability . thus @xmath99 more generally , if we want to consider other variations of the knights model that lack property b , we need only that for any crossing without a t - junction , some local configuration of changes in a bounded region around the crossing can create a t - junction . the induced mapping can be many - to - one , so long as the `` many '' is bounded ( which follows automatically from the restriction that the configuration changes occur in a bounded region around the crossing ) . this will more generally give @xmath100 for some @xmath101 . since the configuration changes only take place within a rectangle @xmath90 , they do not affect whether or not we have @xmath102 for @xmath103 , and we can write the above inequality for a state where we specify whether these other @xmath102 occur . for example , for @xmath104 we might write @xmath105 with the same @xmath106 as above . this can be intuitively thought of as treating the different probabilities of forming t - junctions as independent . repeatedly using equation [ eq : sampleg ] in different subrectangles , we find that @xmath107 we are thus exponentially unlikely to have no t - junctions . in this appendix we prove the existence of the finite structures discussed in the section on jamming percolation in three - dimensions . before giving the formal proof , we sketch the qualitative idea behind the construction of these structures . in the structure for the knights model , shown in figure [ fig : sub1to2 ] , there are two parallelograms consisting of @xmath70 ( northeast - southwest ) chains . the two parallelogragrams are parallel to each other , and are connected and stabilized by @xmath71 ( northwest - southeast ) chains . in two dimensions , such a figure can only be constructed by having some chains cross each other , and this results in a long @xmath70 chain connecting the two ends , unless the model violates property a. however , in three dimensions , the @xmath71 chains can always be run in a direction independent of the plane of the @xmath70 parallelograms , so such a substructure can always be formed , regardless of the neighboring relations . we now formalize this argument . recalling that @xmath108 , @xmath109 , @xmath110 , and @xmath111 are four three - dimensional vectors , with the first three being linearly independent and @xmath111 not equal to any of the first three , there must exist @xmath112 , @xmath113 , @xmath114 , and @xmath46 such that if all of the @xmath119 are positive , then it is easy to make the desired structure . let @xmath120 . then make the structure in figure [ fig : simpleline ] , where each of the vectors @xmath121 represents an a - b chain of length @xmath122 , and @xmath123 represents a c - d chain of length @xmath46 . this structure has the desired properties . if all of the @xmath119 are negative , we simply replace @xmath124 with @xmath125 and use the same argument . we are now left with the most complicated case , where some @xmath119 are positive , and some @xmath119 are negative . in this case , we redefine the @xmath119 , and rewrite equation [ eq : lineardependence ] as where @xmath46 , all @xmath119 , and all @xmath127 , are positive , and for any @xmath90 , either @xmath119 or @xmath127 is zero . then define @xmath128 and @xmath129 . @xmath121 and @xmath130 are both nonzero and linearly independent . we can now make the structure shown in figure [ fig : substructure ] the vectors @xmath131 and @xmath132 represent a - b chains of length @xmath133 and @xmath134 . in these chains every site is stable by the a - b condition except for @xmath83 and @xmath135 . the vectors @xmath136 are chains of length @xmath46 in which every site is stable by the c - d condition . next , since @xmath137 form a complete basis for three - dimensional space , and @xmath138 are only two vectors , there exists a vector @xmath139 , with all @xmath140 , such that @xmath141 is linearly independent of @xmath142 . ( note : this is where the three - dimensional case differs from the two - dimensional case . for models such as the spiral and sandwich models , the two vectors @xmath121 and @xmath130 already span the space . ) we can then make a second copy of figure [ fig : substructure ] , displaced from the original by @xmath141 , as shown in figure [ fig : fullstructure ] . we can now check that in figure [ fig : fullstructure ] all sites except the start and end sites , @xmath83 and @xmath84 are stable under the culling rules , and that @xmath83 and @xmath84 have neighbors from @xmath143 and @xmath144 , respectively . it remains to check that there is no a - b chain connecting @xmath83 to @xmath84 in this structure . there is no obvious such chain , but depending on the vectors @xmath116 and @xmath145 , it is possible that there are some sites in the @xmath121 , @xmath130 , and @xmath141 chains by chance are separated by a vector @xmath116 , inadvertently forming an a - b chain between @xmath83 and @xmath84 . however , if this is the case , we can create new larger structure , simply by multiplying @xmath46 , all @xmath119 , all @xmath127 , and all @xmath146 by the same multiplicative constant . the structure thus grows larger , while the vectors @xmath116 stay the same , so for a sufficiently large multiplicative constant , it is impossible for the different chains in the structure to be adjacent by @xmath116 connections . we thus have a structure with the desired properties . l. berthier , g. biroli , j.p . bouchaud , l. cipelletti , _ et al . _ , science * 310 * , 1797 ( 2005 ) . for earlier hints of a diverging length scale , see n. menon and s. r. nagel , phys . 74 * , 1230 ( 1995 ) . e. r. weeks and d. a. weitz , phys . . lett . * 89 * , 095704 ( 2002 ) . e. r. weeks and d. a. weitz , chem . phys . * 284 * , 361 ( 2002 ) . c. toninelli , g. biroli , and d. s. fisher , phys . lett . * 92 * , 185504 ( 2004 ) . c. toninelli , g. biroli and d. s. fisher , j. stat . phys . * 120 * , 167 ( 2005 ) . g. h. fredrickson and c. h. andersen , phys . lett . * 53 * , 1244 ( 1984 ) . see ref . @xcite , and references therein . c. toninelli and g. biroli , `` jamming percolation and glassy dynamics , '' j. stat . phys . * 126 * , 731 ( 2007 ) . c. toninelli , g. biroli , and d. s. fisher , phys . lett . * 98 * , 129602 ( 2007 ) . m. jeng , j. m. schwarz , phys . rev . lett . * 98 * , 129601 ( 2007 ) . the kosterlitz - thouless transition also exhibits unusual characteristics . there is a discontinity in the spin wave stiffness , or the free energy cost to applying a gradient , accompanied with an exponentially diverging correlation length . however , there is no magnetization , or local order parameter , that goes to zero at the transition . the jamming percolation models have a simple order parameter , the fraction of sites in the infinite cluster . j. chalupa , p. l. leath , and g. r. reich , j. phys . c * 12 * , l31 ( 1979 ) . b. pittel , j. spencer , and n. wormald , j. comb . series b * 67 * , 111 ( 1996 ) . p. m. kogut , j. phys . c : solid state phys . * 14 * , 3187 ( 1981 ) . n. s. branco , s. l. a. de queiroz , and r. r. dos santos , j. phys . c : solid state phys . * 19 * , 1909 ( 1986 ) . j. adler and d. stauffer , j. phys . a. : math . gen . * 23 * , l1119 ( 1990 ) . a. c. d. van enter , j. stat . phys . * 48 * , 943 ( 1987 ) . r. h. schonmann , ann . prob . * 20 * , 174 ( 1992 ) . m. aizenman and j. l. lebowitz , j. phys . a. * 21 * , 3801 ( 1988 ) . j. adler , physica a * 171 * , 453 ( 1991 ) . s. leonard et . , `` non - equilibrium dynamics of spin facilitated glass models , '' cond - mat/0703164 , to appear in j. stat . mech . : theory and expt . c. s. ohern , s. a. langer , a. j. liu , and s. r. nagel , phys 88 * , 075507 ( 2002 ) . h. hinrichsen , adv . phys . * 49 * , 815 ( 2000 ) . this observation was kindly pointed by c. toninelli and g. biroli in arxiv:0709.0583 after an initial draft of our paper was made available making the opposite claim .	we investigate kinetically constrained models of glassy transitions , and determine which model characteristics are crucial in allowing a rigorous proof that such models have discontinuous transitions with faster than power law diverging length and time scales . the models we investigate have constraints similar to that of the knights model , introduced by toninelli , biroli , and fisher ( tbf ) , but differing neighbor relations . we find that such knights - like models , otherwise known as models of jamming percolation , need a `` no parallel crossing '' rule for the tbf proof of a glassy transition to be valid . furthermore , most knight - like models fail a `` no perpendicular crossing '' requirement , and thus need modification to be made rigorous . we also show how the `` no parallel crossing '' requirement can be used to evaluate the provable glassiness of other correlated percolation models , by looking at models with more stable directions than the knights model . finally , we show that the tbf proof does not generalize in any straightforward fashion for three - dimensional versions of the knights - like models .
hadronic transitions between states of the available heavy quarkonia , charmonium and bottomonium , present a very interesting case study in dynamics of both the heavy quarks and the light mesons emitted in the transitions . in particular , the observed properties of such processes with the emission of two pions are understood within the chiral low - energy dynamics of pions , starting with the earliest observation @xcite of the decay @xmath3 and the theoretical analyses of the data @xcite . furthermore , the qcd picture of such decays is that the heavy quarkonium transition generates soft gluonic field which then produces the light mesons . the heavy quarkonium can be considered as a compact and nonrelativistic object in its interaction with a soft gluonic field , which justifies the use of the multipole expansion in qcd in analysing these processes @xcite . in this approach the amplitude of the process factorizes into the heavy quarkonium part , i.e. the transition between the levels as a source of the field , and the light meson part , which describes the creation of the mesons by the field operators . the former part depends on the dynamics of the quarkonium , while the latter one can be understood in some detail by combining @xcite the methods of chiral dynamics and the general low - energy relations in qcd , in particular those for the anomalies in the trace of the energy - momentum tensor and in the singlet axial current . each of these two factors in the amplitudes of the discussed hadronic transitions can be used , in certain extent , independently . for instance , the decays @xmath4 and @xmath5 ( as well as @xmath6 and @xmath7 ) have essentially the same heavy quarkonium part , so that the ratio of the decay rates is fully determined by the low - energy meson amplitudes @xcite . on the other hand the charmonium amplitude for the double interaction with the chromo - electric field ( the chromo - polarizability ) , which dominates the quarkonium factor in the decay @xmath3 can be used @xcite for considering the low - energy limit of the conversion of @xmath8 into @xmath9 on a nucleon : @xmath10 , where the soft gluonic field is that inside the nucleon . the same charmonium transition amplitude can be used as a benchmark for the diagonal chromo - polarizability of @xmath9 , which determines the low - energy elastic scattering of @xmath9 on a nucleon @xcite . for these reasons a more detailed understanding of both the heavy quarkonium amplitudes and of those determining the gluon conversion into light mesons is desirable . presently a quite detailed experimental data are available on the dominant hadronic transitions in charmonium @xcite : @xmath3 , and in bottomonium @xcite : @xmath6 and @xmath11 . numerous other , less visible , transitions have been observed , e.g. @xmath12 @xcite . thus it is not unreasonable to expect , given the capabilities of the modern @xmath13 experiments at the charm and bottom thresholds , that a significantly improved data will become available . such data may enable a study of finer effects in the amplitudes of the hadronic transitions beyond the leading ones thus far considered in the literature . the purpose of the present paper is to analyze sub - leading relativistic effects in the amplitudes of the two - pion transitions in charmonium and bottomonium , and possible ways of studying such effects from the data . for the transitions between the @xmath14 states of quarkonium , dominated by the second order in the leading @xmath15 term of the multipole expansion , these corrections arise from the second order in the @xmath16 interaction . it will be argued that in the charmonium transition @xmath3 the correction adds coherently to the leading term as an @xmath17 term . being of a rather moderate value in this particular decay , the correction is enhanced by a factor of 3 in the transition @xmath18 with a potentially very noticeable effect in the spectrum of the dipion invariant mass . furthermore , this correction modifies the estimate of the charmonium chromo - polarizability with implications for charmonium scattering on nucleons . it will be further argued that , although generally the relativistic effects in bottomonium are quite small , in the particular case of the decay @xmath19 it appears that the leading nonrelativistic term it strongly suppressed dynamically , and including the relativistic effects of the @xmath16 interaction and of the @xmath20 mixing can possibly resolve the long standing puzzle of the peculiar observed spectrum of the dipion invariant masses in this decay , which spectrum by far does not conform with the predictions from the leading chromo - electric interaction . also some consideration will be given to the significance of the final state interaction ( fsi ) between the pions in the two - pion transitions . the fsi effects have been discussed ever since the earliest analyses @xcite , and are generally believed to be moderate in the kinematical region of the decays @xmath21 and @xmath6 . however once the amplitudes of the transitions are studied in finer detail , one will have to better quantify these effects . although at present the fsi effects can not be fully analyzed , their behavior at low invariant mass of the two pions can be estimated within the chiral expansion . for this purpose we give here a systematic derivation of the amplitudes for two pion production by gluonic operators in the leading chiral limit , including the terms with the pion mass , and consider the first sub - leading term due to iteration of the chiral amplitudes . at the invariant mass of the two pions near the threshold these corrections are quite small - only few percent . their continuation to higher values of the invariant mass can be studied experimentally in a rather straightforward way . such study , in particular , would quantify the impact of the fsi effects on the estimates of the quarkonium amplitudes from the observed decay rates . it will be also argued that even with the already available data such impact can be estimated as quite moderate , as generally expected . the material in the present paper is organized as follows . in section 2 the chromo - electric and chromo - magnetic dipole interactions responsible for the two - pion transitions are discussed and the parameters determining the relevant quarkonium amplitudes to order @xmath22 are introduced . the section 3 contains a compilative derivation in the leading chiral order of the amplitudes for creation of the pion pair by gluonic operators , and in section 4 the discussion of the two previous sections is combined in the expressions for the two - pion transition amplitudes with the leading relativistic terms included . the latter expressions are confronted with the available date on the transitions @xmath23 and @xmath24 in section 5 and the significance of the chromo - magnetic interaction in transitions between charmonium resonances is evaluated . in section 6 the fsi effects are considered using iteration of the chiral amplitudes and it is also argued on phenomenological grounds that these effects are only of a moderate magnitude in the observed two - pion transitions . the section 7 addresses the long - standing puzzle of the unusual spectrum of the dipion invariant masses observed in the decay @xmath25 and it is suggested that this spectrum can be explained if the formally leading nonrelativistic quarkonium amplitude is dynamically suppressed in this transition , so that the relativistic terms provide a significant contribution . finally , section 8 contains a summary of the discussion in the present paper . considering quarkonium as a compact nonrelativistic object one can apply the multipole expansion for the quarkonium interaction with soft gluonic field @xcite . the leading term in this expansion is the chromo - electric dipole interaction with the chromo - electric field @xmath26 described by the hamiltonian h_e1=-1 2 ^a r ^a ( 0 ) , [ e1 ] where @xmath27 is the difference of the color generators acting on the quark and antiquark ( e.g. @xmath28 with @xmath29 being the gell - mann matrices ) , and @xmath30 is the vector for relative position of the quark and the antiquark . the convention used throughout this paper is that the qcd coupling @xmath31 is included in normalization of the gluon field operators , so that e.g. the gluon field lagrangian is written as @xmath32 . the amplitude of the two - pion transition between heavy quarkonium states @xmath33 ( with @xmath34 and @xmath35 used as a generic notation for the initial and the final quarkonium states ) can thus be written as are considered here , since the emission of neutral pions @xmath36 is trivially related by the isospin symmetry . ] _ 1 ) = 1 2 ^+ ^- \| e^a_i e^a_j \| 0 ^(12)_i j , [ e1ampg ] where @xmath37 can be termed , in complete analogy with the atomic properties in electric field , as the transitional chromo - polarizability of the quarkonium . in other words , the @xmath38 transition in the chromo - electric field is described by the effective hamiltonian h_eff=-1 2 ^(12)_i j e^a_i e^a_j , [ heff ] with the chromo - polarizability given by ^(12)_i j=1 16 _ 1 \| ^a r_i g r_j ^a \| _ 2 , [ aij ] where @xmath39 is the green s function of the heavy quark pair in a color octet state . the latter function is not well understood presently , so that an _ ab initio _ calculation of the chromo - polarizability would be at least highly model dependent . also a simple remark is in order that the discussed decays in quarkonia are governed by transitional polarizabilities , i.e. those linking different states of quarkonium . the diagonal chromo - polarizability of quarkonium states , in particular of the charmonium resonances @xmath8 and @xmath9 can also be measured @xcite and is relevant to the problem of scattering of these resonances on nuclei @xcite . generally @xmath40 is a symmetric tensor . clearly , for transitions between pure @xmath41 wave quarkonium states this tensor is necessarily proportional to @xmath42 : @xmath43 . however for the @xmath14 states a @xmath20 mixing generally takes place due to the relativistic effects in the order @xmath22 so that a traceless @xmath44 wave part of the chromo - polarizability is also present and is proportional to spin-2 combination of the polarizations , @xmath45 and @xmath46 of the initial and the final states : _ i j = _ 0 _ ij ( _ 1 _ 2 ) + _ 2 . [ asd ] numerically , the @xmath20 mixing in charmonium appears to be comparable with the characteristic value of the relativistic parameter for this system , @xmath47 . in particular , rosner @xcite using the @xmath13 decay width of @xmath48 and considering this resonance as dominantly a @xmath49 state , estimates the angle @xmath50 of the @xmath51 mixing , i.e. the @xmath20 mixing for @xmath8 , as @xmath52 . it can be noticed however that the mixing does not contribute to the spinless part @xmath53 of the chromo - polarizability and that in the spin - averaged transition rate there is no interference between the @xmath41 and @xmath44 wave parts of @xmath54 . thus the effect of the @xmath20 mixing in the rate is of order @xmath55 and is thus generally very small . it should be also mentioned that the effects of the recoil of the final quarkonium state in the decay also arise in the order @xmath22 and contribute terms quadratic in the momentum of the pion pair @xmath56 to the spin - diagonal part of the tensor polarizability @xmath54 . however such terms would also be of higher order in the chiral expansion in the pion momenta , which is used here . therefore the recoil terms are suppressed by a product of two small parameters and can be neglected within the approximations adopted in this paper . besides the discussed @xmath20 mixing , the only other effect in the order @xmath22 in the amplitudes of two - pion transitions between quarkonium @xmath41 wave states arises through the second order in the m1 interaction with the chromo - magnetic field @xmath57 described by the hamiltonian h_m1= - 1 2 m ^a ( ^a ) , [ m1 ] where @xmath58 is the difference of the spin operators acting on the quark and the antiquark , and @xmath59 is the heavy quark mass . the contribution of this term to the amplitude of two - pion transition between the @xmath41 states of quarkonium can be written as @xmath60~,\end{aligned}\ ] ] for the transitions between @xmath14 states , and a_m(_2 ^+ ^- _ 1 ) = 3 2 ^+^-\| b_i^a b_i^a \|0 _ m^(12 ) [ am0 ] for the transitions between @xmath61 states . ( the notation @xmath62 and @xmath63 is used here for the states of quarkonium in order to emphasize that this relation is specific for the @xmath61 states . ) the amplitude @xmath64 is a chromo - magnetic analog of the chromo - electric term @xmath54 and is formally given by the formula _ m^(12)=1 48 m^2 _ 1 \| ^a g ^a \| _ 2 , [ amg ] where @xmath65 and @xmath66 are the coordinate parts of the @xmath41-wave wave functions of the initial and the final states . it is taken into account here that the amplitudes generated by the m1 interaction are already suppressed in comparison with those described by eq.([e1ampg ] ) by the factor @xmath22 . therefore the wave functions of the quarkonium states can be taken in the form where the spin and coordinate parts are factorized . the spin-2 part of the amplitude in eq.([am1 ] ) contributes to the decay rate only in the order @xmath55 , similarly to the effect of the @xmath67 mixing . however , the spin-0 part in this amplitude as well as that in eq.([am0 ] ) ( the only one present there ) does interfere with the leading nonrelativistic amplitude in eq.([e1ampg ] ) proportional to @xmath53 from eq.([asd ] ) . therefore this part provides the first relativistic correction , proportional to @xmath22 , to the transition rate . one can also notice that the effect for the @xmath61 states is three times bigger than for the @xmath14 states . as one can see from the previous section , a crucial role in the discussed approach to calculating the two - pion transition amplitudes is played by the amplitudes for production of two pions by operators quadratic in components of the gluon field strength tensor . therefore in this section we consider the general amplitude of such type : @xmath68 , describing the creation of two pions by the local gluonic operator @xmath69 . in the leading chiral limit the momenta @xmath70 and @xmath71 of the pions as well as the pion mass @xmath72 are to be considered as small parameters , and the expression for the amplitude , quadratic in these parameters , can be written in the following general form - ^+(p_1 ) ^-(p_2)\| f_^a f_^a\|0 = ( g _ g _ - g _ g_)+ z t _ , [ genf ] where the structure @xmath73 has zero overall trace : @xmath74 , and @xmath75 , @xmath76 , and @xmath77 are yet undetermined coefficients . the form of the amplitude in eq.([genf ] ) is uniquely determined by the symmetry ( with respect to the indices ) of the operator @xmath78 and by the adler zero condition , which requires that the amplitude goes to zero when either one of the two pion momenta is set to zero and the other one is on the mass shell . ] . the coefficients @xmath75 and @xmath76 in eq.([genf ] ) are in fact determined @xcite by the anomaly in the trace of the energy - momentum tensor @xmath79 in qcd . indeed , in the low - energy limit in qcd , i.e. in qcd with three light quarks , one finds _ ^=-b 32 ^2 f^a _ f^a + _ q = u , d , s m_q ( \|q q ) , [ anom ] where @xmath80 is the first coefficient in the beta function for qcd with three quark flavors . the first term in eq.([anom ] ) represents the conformal anomaly , while the quark mass term arises due to the explicit breaking of the scale invariance by quark masses . on the other hand , the matrix element of the energy - momentum tensor @xmath79 over the pions : @xmath81 is fully determined @xcite in the quadratic in @xmath82 and @xmath72 order by the conditions of symmetry in @xmath83 and @xmath84 , conservation on the mass shell ( @xmath85 at @xmath86 ) , normalization ( @xmath87 at @xmath88 ) , and the adler zero condition ( @xmath89 ) : _ ( p_1,p_2)= g _ -p_1 p_2- p_1 p_2 . [ st ] the equations ( [ genf ] ) and ( [ st ] ) allow for the pion momenta to be off - shell in order to demonstrate the adler zero . however in what follows only the amplitudes with pions on the mass shell will be considered , so that it will be henceforth implied that @xmath86 . in particular one finds for the trace of the expression in eq.([st ] ) _ ^(p_1,p_2)=2(p_1 p_2 ) + 4m^2 . [ tst ] furthermore , the quark mass term in eq.([anom ] ) , corresponding to the explicit breaking of the chiral symmetry in qcd corresponds to the same symmetry breaking by the pion mass term in the pion theory . thus one finds to the quadratic order in @xmath90 : ^+ ^- \|_q = u , d m_q ( \|q q)\| 0 = m^2 , [ mt ] while the term with the strange quark makes no contribution to the discussed amplitude . combining the formula in eq.([anom ] ) for @xmath91 with the expressions ( [ tst ] ) and ( [ mt ] ) one finds the matrix element of the gluonic operator over the pions in the form - ^+(p_1 ) ^-(p_2)\| f_^a f^a \|0 = 32 ^2 b [ gt ] which thus allows to determine the coefficients @xmath75 and @xmath76 in eq.([genf ] ) : @xmath92 and @xmath93 . the coefficient @xmath77 of the traceless part in eq.([genf ] ) can not be found from the trace relation ( [ anom ] ) . novikov and shifman @xcite estimated this coefficient by relating this part to the matrix element of the traceless ( twist - two ) energy - momentum tensor of the gluons only : @xmath94 , z t_^= 4 _ s ^+(p_1 ) ^-(p_2)\|_^g \|0 . [ nsr ] they then assume that the matrix element of the twist - two operator is proportional to the traceless part of the phenomenological energy momentum tensor of the pions , ^+(p_1 ) ^-(p_2)\|_^g \|0 = _ g [ rhog ] with the proportionality coefficient interpreted , similarly to the deep inelastic scattering , as the fraction of the pion momentum carried by gluons " . they further introduce a related parameter @xmath95 and conjecture that numerically @xmath96 . for the purpose of the present consideration the interpretation of the parameter @xmath97 as being related to the pion gluon structure function is not essential and we treat @xmath97 here as a phenomenological and measurable , and actually measured , parameter . summarizing the results so far in this section one can write the expression for the general matrix element ( [ genf ] ) for on - shell pions as - ^+(p_1 ) ^-(p_2)\| f_^a f_^a\|0 = 8 ^2 3 b , [ kapf ] where @xmath98 is the total four - momentum of the dipion . few remarks are due regarding effects of higher order in @xmath99 . the trace term in eq.([kapf ] ) receives no renormalization , provided that the coefficient @xmath100 is replaced by @xmath101 with @xmath102 being the full beta function in qcd . this modification however only affects the overall normalization of the trace part , and can in fact be absorbed into the definition of the heavy quarkonium amplitudes . on the contrary , the relative coefficient of the traceless term in eq.([kapf ] ) , i.e. the parameter @xmath97 , does depend on the normalization scale , which scale is appropriate to be chosen as the characteristic size of the heavy quarkonium @xcite . however , given other uncertainties in the analysis of the two - pion transitions , the logarithmic variation of @xmath97 is a small effect . in particular , this effect is likely to be smaller than the discussed in this paper relativistic effects in the amplitudes of the two - pion emission . the matrix element in eq.([kapf ] ) describes the production of the two pions in two partial waves in their center of mass system : the @xmath41 wave and the @xmath44 wave . the two waves can be measured separately , and also any effects of the final state interaction between pions are different in these two orbital states . therefore it is quite instructive for the subsequent discussion to explicitly separate the @xmath41 and @xmath44 waves in the matrix element , i.e. to rewrite the amplitude ( [ kapf ] ) in the form - ^+(p_1 ) ^-(p_2)\| f_^a f_^a\|0 = s_+ d _ . [ sd ] clearly , the trace term in eq.([kapf ] ) corresponds to a pure @xmath41 wave , while the traceless term proportional to @xmath97 contains both waves . in order to perform explicit partial wave separation in @xmath103 it is helpful to introduce @xcite the four vector @xmath104 describing the relative momentum of the two pions , which reduces to a purely spatial vector in the c.m . system of the pions ( @xmath105 ) . then the tensor _ = r_r_+ 1 3 ( 1- 4 m^2 q^2 ) ( q^2 g_- q_q _ ) [ lmn ] is also purely spatial in the c.m . frame and corresponds to pure @xmath44 wave . using this tensor one can make the following series of replacements for the terms of the generic structure @xmath106 in the tensor @xmath107 : @xmath108 here in the first replacement the cross terms between @xmath109 and @xmath110 are dropped since they cancel in @xmath103 due to the c symmetry ( @xmath111 ) , while the @xmath112 term in the last transition is dropped , since such structure cancels in the traceless tensor @xmath113 . using eq.([ppsub ] ) one readily finds from the formula ( [ kapf ] ) the expressions for the @xmath41 and @xmath44 wave amplitudes : @xmath114 \right \}~,\end{aligned}\ ] ] and d_=8 ^2 3 b 9 4 ( _ g_+ _ g_- _ g_- _ g _ ) . using the formulas in the equations ( [ e1ampg ] ) , ( [ asd ] ) and ( [ am1 ] ) and the expressions ( [ sw ] ) and ( [ dw ] ) for the dipion production amplitudes in the chiral limit , one can readily find the amplitude of the transition @xmath115 between generic @xmath14 states of a heavy quarkonium . after a straightforward calculation one finds the @xmath41 wave decay amplitude @xmath116 \,(\psi_1 \cdot \psi_2)~,\end{aligned}\ ] ] as well as three types of @xmath44 wave amplitude : one unrelated to the spins of the quarkonium states d_1(_2 ^+ ^- _ 1)=-4 ^2 b _ 0^(12 ) ( 1+_m ) 3 2 _ p^p^p^2 ( _ 1 _ 2 ) , [ d1 ] and two amplitudes with the correlation with the polarization of the initial and the final resonances d_2(_2 ^+ ^- _ 1)= 4 ^2 b _ 0^(12 ) ( _ 2 + 3 2 _ m ) 2 ( 1 + 2 m^2 q^2 ) q_q_^ [ d2 ] and d_3(_2 ^+ ^- _ 1)= 4 ^2 b _ 0^(12 ) ( _ 2 + 3 2 _ m ) 3 4 _ [ d3 ] in these formulas the following notation is used : @xmath117 stands for the 4-momentum of the initial quarkonium resonance , @xmath118 and @xmath119 are the polarization 4-vectors for the @xmath14 states , and @xmath120 is the spin-2 structure @xmath121 . finally , @xmath122 and @xmath123 stand for the ratia _ m = _ m _ 0 , _ 2 = _ 2 _ 0 [ chis ] and encode the relative magnitude of the @xmath124 relativistic effects due to respectively the chromo - magnetic interaction ( eq.([m1 ] ) ) and the @xmath125 mixing . the three @xmath44 waves correspond to different angular correlations . the first one , @xmath126 , given by eq.([d1 ] ) corresponds to a @xmath44-wave motion in the c.m . frame of two pions , which correlates with the motion of the c.m . system in the laboratory frame , i.e. with the direction of @xmath56 . this @xmath44 wave arises in the leading nonrelativistic approximation @xcite and is in fact observed and measured experimentally @xcite for the transition @xmath3 . the second @xmath44-wave amplitude , @xmath127 in eq.([d2 ] ) , corresponds to the two pions being in the @xmath41 wave in their c.m . system and describes the correlation of the spins of the initial and the final resonances with the @xmath44-wave motion of the two - pion system as a whole . finally , the amplitude @xmath128 given by eq.([d3 ] ) corresponds to a @xmath44-wave motion of the pions in their c.m . frame , which @xmath44 wave is correlated with the spins of the quarkonium states . it can be noted that the two latter amplitudes are proportional to a product of two relatively small parameters @xmath97 and @xmath129 . neither @xmath127 nor @xmath128 have yet been observed experimentally , although a study @xcite of polarization effects in the decay @xmath130 , utilizing a transversal polarization of the doris beams qualitatively confirms that these spin - dependent amplitudes are quite small . ( a discussion can be found in the review @xcite . ) the transitions between @xmath61 states of quarkonium have not been observed yet . one may hope however that with a dedicated effort a two - pion transition from the recently found @xmath131 resonance : @xmath18 can be observed and studied . within the approach discussed here such transition is closely related to the familiar decay @xmath3 , and in fact can be used for a useful calibration of the total width of @xmath131 @xcite . clearly , on the theoretical side the transitions between @xmath61 states are simpler than those between the @xmath14 ones since no polarization effects are involved . on the other hand the effect of the m1 interaction ( eq.([m1 ] ) ) is enhanced for the @xmath61 states ( eq.([am1 ] ) ) by a factor of 3 , so that the relevant transition amplitudes of a generic @xmath132 transition are given by @xmath133~,\end{aligned}\ ] ] and d_1(_2 ^+ ^- _ 1)= 4 ^2 b _ 0^(12 ) ( 1 + 3 _ m ) 3 2 _ p^p^p^2 . [ dep ] the implications of the enhancement of the relativistic term for the spectrum of the dipion invariant masses is discussed in the next section . the most experimentally studied so far transitions of the discussed type are @xmath3 and @xmath6 . in particular , in the bes experiment @xcite both the spectrum of the invariant masses of the two - pion system and the angular distribution described by the @xmath44 wave amplitude of the type @xmath126 ( eq.([d1 ] ) were analyzed , using the formulas equivalent to eq.([spp ] ) and eq.([d1 ] ) with @xmath134 , and the parameter @xmath97 was determined from the fits . the fit to the mass spectrum resulted in the value @xmath135 , while the fit to the ratio of the @xmath126 and @xmath41 wave amplitude gave @xmath136 . clearly , the consistency of these two values implies that the discussed approach correctly predicts @xcite the ratio of the @xmath126 wave in terms of the sub - dominant term proportional to @xmath97 in the @xmath41 wave amplitude . furthermore , considering the modification of the expressions for these amplitudes due to the relativistic parameter @xmath122 , it is clear that the fit parameter in the experimental analysis is not exactly @xmath97 but rather _ eff^( ) = 1+_m^( ) 1-_m^( ) ( 1 + 2 _ m^( ) ) . [ kappaeff ] in the estimates in this section we use the value with the smaller error : @xmath137 . a similar analysis @xcite of the decay @xmath6 was inconclusive regarding the @xmath126 wave amplitude , whereas the fit for the dipion invariant mass distribution resulted in @xmath138 . thus the data display a statistically significant decrease of the parameter @xmath139 in bottomonium in comparison with the similar parameter in charmonium : @xmath140 . some decrease of this type of the parameter @xmath97 was in fact predicted @xcite based on the different characteristic scale for normalization of the gluonic operator in eq.([kapf ] ) relevant for the transitions in the two systems . indeed , the characteristic size of @xmath141 , @xmath142 , is somewhat smaller than that of @xmath9 , @xmath143 . however this effect is rather difficult to quantify , given that at both these scales the applicability of the perturbation theory in qcd is marginal . considering also that the difference in the size is not large enough , especially in the logarithmic scale , in order to explain the observed difference in the values of @xmath139 , one may explore another approach to the observed difference by neglecting the variation of the parameter @xmath97 altogether and ascribing the observed effect to different values of the relativistic term @xmath122 in these two quarkonium systems . indeed , the relativistic parameter @xmath22 in the lowest states of bottomonium is only about one third of that in charmonium , as can be inferred e.g. by comparing the excitation energies relative to the mass : @xmath144 . thus one can also expect @xmath145 , and assuming that the value of @xmath97 in eq.([kapf ] ) is the same for the transitions in both quarkonia , one can estimate the relativistic effect of the m1 interaction from -1 2 _ m^( ) - 2 _ m^( ) 0.27 0.07 . [ kest ] taken at face value , such estimate gives @xmath146 if @xmath147 is set equal to @xmath148 and in @xmath149 if @xmath147 is neglected altogether . in either case the difference between the specific numerical estimates is within the experimental and theoretical uncertainty , while the estimated value compares well with the expected magnitude of the relativistic effects in charmonium . a more definite determination of the parameter @xmath150 could be enabled by observing the distribution in the dipion mass in the decay @xmath18 . indeed , as previously mentioned , the effect of the chromo - magnetic m1 interaction is enhanced in this decay by a factor of 3 . if the value of @xmath122 for the charmonium transitions is in the estimated range @xmath151 , the effective parameter @xmath152 for the latter decay can amount to 0.3 - 0.45 . although the prospect of measuring the @xmath126 wave in the decay is likely very remote , the effect of such larger value of @xmath139 should be visible in the more experimentally accessible spectrum of the dipion mass , as illustrated in fig.1 . as can be seen from the plots the larger value of @xmath152 results in a significant suppression of the spectrum at low invariant mass , due to the zero of the @xmath41 wave amplitude at a higher , than for @xmath3 , value of @xmath153 . also the total rate contains an overall suppression due to the factor @xmath154 vs. the factor @xmath155 for the transition between the @xmath14 states . so far the amplitudes of the two - pion transitions were considered here in the chiral limit . the formulas in eq.([kapf ] ) and in eqs.([sw ] ) and ( [ dw ] ) are exact in the leading chiral order , i.e. as far as the quadratic terms in the pion momenta and mass are concerned . the only dynamical modification of these expressions can arise from the previously mentioned qcd renormalization effects . in particular these expressions get no corrections due to the final state interaction ( fsi ) between the pions . the latter interaction however can give rise to the terms whose expansion starts with the fourth power of momenta and the pion mass , and generally can modify the amplitude at momenta of the pions relevant for actual transitions in quarkonium . the effects of fsi in chiral treatment of the two - pion transitions in quarkonium were a matter of concern ever since the earlier theoretical analyses @xcite . the effect in the phases of the amplitudes is well known : these phases for the production amplitudes are equal to the two - pion scattering phases in the corresponding partial waves : @xmath156 , @xmath157 , where the @xmath158 phases for the @xmath41 wave , @xmath159 , and for the @xmath44 wave , @xmath160 are quite well studied . decay does not take into account the relative phase between the @xmath41 and @xmath44 wave pion production amplitudes . thus it would be interesting to know whether including the phase factor in the angular analysis , produces a significant impact on the results . ] it is also generally estimated both on theoretical and phenomenological grounds that the fsi corrections are not big ( at most 20 - 25% ) in the transitions @xmath161 and @xmath24 . ( for a discussion see the review @xcite . ) some phenomenological arguments in favor of such estimate will also be discussed further towards the end of this section , and we start with a theoretical estimate of the onset of the higher term in the chiral expansion . the interaction of pions at low energy in the @xmath44 wave is quite weak , so that any modification by fsi of the @xmath44 wave production amplitude of eq.([dw ] ) can be safely neglected , and only the modification of the @xmath41 wave amplitude ( [ sw ] ) , @xmath162 , is of interest for present phenomenology . the imaginary part of the correction at @xmath163 is found from the unitarity relation in terms of the isospin i=0 @xmath164 scattering amplitude @xmath165 in the @xmath41 wave as ( s _ ) = t(q^2 ) 16 s _ . [ unit ] the amplitude @xmath165 is well known in the chiral limit , i.e. in the quadratic in @xmath110 and @xmath72 approximation , since the work of weinberg @xcite . in the normalization used here this amplitude has the form t(q^2)=2 q^2 - m^2 f_^2 , [ tpipi ] where @xmath166mev is the @xmath167 decay constant . clearly , the expression in eq.([unit ] ) is of the fourth power in @xmath110 and @xmath72 . the real part of @xmath162 can then be estimated from eq.([unit ] ) using the dispersion relation in @xmath153 for the amplitude @xmath41 . in doing so one should set the condition for the subtraction constants that this real part does not contain quadratic ( and certainly also constant ) terms in @xmath110 and @xmath72 , since these are given by eq.([sw ] ) . after these subtractions the remaining dispersion integral is still logarithmically divergent and contains the well known ` chiral logarithm ' , depending on the ultraviolet cutoff @xmath168 , which is usually set at @xmath169gev , i.e. the scale where any chiral expansion certainly breaks down . using the equations ( [ sw ] ) , ( [ unit ] ) and ( [ tpipi ] ) one can readily estimate the first fsi correction with a ` logarithmic accuracy ' . the expression for the full @xmath41 wave production amplitude then takes the form @xmath170 \right \}~,\end{aligned}\ ] ] where the correction terms @xmath171 and @xmath160 to eq.([sw ] ) are given as _ 1 = 2 ( q^2)^2- 7 q^2 m^2 + 3 m^4 16 ^2 f_^2 ( q^2+m^2 ) + i 2 q^2 - m^2 16 f_^2 , [ xi1 ] and _ 2 = 2 ( q^2)^2- 9 q^2 m^2 + 8 m^4 16 ^2 f_^2 ( q^2 + 2 m^2 ) + i 2 q^2 - m^2 16 f_^2 , [ xi2 ] where the non - logarithmic imaginary part is retained for reference regarding the normalization . the lower limit under the logarithm is generally a function of both @xmath153 and @xmath90 , however any difference of this function from the value @xmath90 used in eqs.([xi1 ] ) and ( [ xi2 ] ) is a non - logarithmic term , i.e. beyond the accuracy of these equations . since @xmath90 is the smallest of the two parameters in the physical region of pion production , it can be expected that using this parameter provides a conservative estimate of the effect of fsi . estimating the corrections in eq.([xi1 ] ) and eq.([xi2 ] ) , one finds that at the lower end of the physical phase space , i.e. near @xmath172 , these terms do not exceed few percent . thus the corrections only weakly modify the normalization of the pion production amplitude near the threshold . a theoretical extrapolation to higher values of @xmath153 is problematic , and , most likely , one would have to resort to using actual data on the dipion spectra in order to judge upon the significance of deviation from the essentially linear in @xmath153 behavior of the amplitude described by eq.([spp ] ) . a quantitative estimate of the deviation from this behavior has been attempted @xcite using the data @xcite on @xmath173 by parametrizing the deviation as a factor @xmath174 in the amplitude with @xmath59 being a parameter . the thus obtained lower limit on @xmath59 is @xmath175gev at 90% c.l . another phenomenological argument in favor of a relatively moderate fsi effect in the absolute value of the dominant @xmath41 wave in the decay @xmath4 stems from the previously mentioned agreement of the observed @xcite value of the ratio @xmath176 with the parameter @xmath97 entering the expression for the @xmath41 wave and extracted from the two - pion invariant mass spectrum . clearly such an agreement would be ruined if there was a significant enhancement of the @xmath41 wave by fsi . furthermore , the observed ratio of the rates of the transitions @xmath177 and @xmath3 reasonably agrees with the calculation @xcite , which neglects any fsi effects in the latter decay , while it is clear that there is no such effects in the emission of one - particle , i.e. of @xmath178 . the expected accuracy of the theoretical calculation is mostly that of applying the nonrelativistic limit to charmonium , i.e. @xmath179 in the amplitude . thus the theoretical ( and experimental ) uncertainty may allow for a presence of the fsi effects in the two - pion system at such level , however large effects of this type , like those recently claimed in ref . @xcite , are definitely excluded . finally , one can notice in connection with the equations ( [ d2 ] ) and ( [ d3 ] ) for the @xmath127 and @xmath128 waves involving the polarizations of the quarkonium resonances , that the unknown quarkonium matrix elements all cancel in the ratio @xmath180 . the @xmath127 amplitude is determined by the @xmath41 wave of the pions in their c.m . system , and the @xmath128 contains the pions in the @xmath44 wave relative to each other . thus if the ratio of actual amplitudes @xmath180 could be measured from angular distributions , e.g. in the decay @xmath4 , this would produce direct data on the modification by fsi of the @xmath41 wave relative to the @xmath44 wave . the decay @xmath25 is known to be quite different from the well behaved " transitions @xmath3 and @xmath6 in that the spectrum @xcite of the dipion invariant masses in this decay has two distinct maxima at low and high values of @xmath153 . the proposed solutions to the puzzle presented by this behavior included presence of a hypothetical exotic resonance @xcite , breakdown of the multipole expansion @xcite , and unusual fsi effects @xcite . in this section i exercise an alternative , and somewhat more conventional explanation of the observed behavior of the spectrum in terms of the considered here terms of the multipole expansion . namely , one can readily verify that the dipion mass spectrum in the decay @xmath25 is reasonably reproduced by the equations ( [ spp ] ) - ( [ d3 ] ) if the parameters @xmath122 and @xmath123 for this decay are rather large : of order one , which is probably a result of a dynamical suppression of the leading nonrelativistic quarkonium amplitude @xmath53 . the overall amplitude of the decay @xmath25 is arguably strongly suppressed . indeed , the observed absolute rate of the transition @xmath181 is only about 0.2 of the rate of @xmath6 in spite of a significantly larger available phase space . the estimated suppression of the former decay would be even stronger , if its amplitude was `` well behaved '' , i.e. given by leading e1 interaction and thus dominantly proportional to @xmath153 . furthermore , the total rate of another similar transition from the same @xmath182 resonance , @xmath183 , is about 0.6 of the rate of @xmath184 , even though the energy released in the former transition is only slightly above the two - pion threshold , and the transition is very strongly kinematically suppressed . thus it appears quite reasonable to conclude that the pure @xmath41 wave part of the quarkonium amplitude in eq.([aij ] ) , @xmath53 is small for the discussed transition . this can be a result of cancellation in the ( appropriately weighed ) overlap of the wave functions determining this amplitude due to the oscillations of the @xmath185 wave function . such cancellation however is not necessarily present in the overlap terms due to the @xmath67 mixing , @xmath186 in eq.([asd ] ) , and due to the m1 interaction , @xmath187 in eq.([amg ] ) . the latter terms are naturally expected to be of the order @xmath124 in comparison with an unsuppressed leading amplitude , e.g. the amplitudes of the well behaved " transitions : @xmath24 or @xmath188 . in other words , the ratia @xmath122 and @xmath123 ( eq.([chis ] ) ) can be large due to the small denominator @xmath53 . a fully quantitative analysis of the decay @xmath25 is complicated by that the energy released in the transition @xmath189mev is rather large for a straightforward application of the chiral - limit formulas for the pions . however , for at least a qualitative estimate , i neglect here this complication and apply the equations ( [ spp ] ) - ( [ d3 ] ) to evaluate the dipion mass spectrum . the resulting behavior is illustrated in fig.2 , where the parameters are set as @xmath190 , @xmath191 , and @xmath192 . clearly , the evaluated dipion mass spectrum closely resembles the experimentally observed @xcite , although no attempt is made here at a quantitative fit to experimental data . in order to assess whether the values of the relativistic parameters @xmath122 and @xmath123 used in the plots of fig.2 are of the order of the expected relativistic effects in bottomonium , it is instructive to compare the corresponding amplitudes @xmath193 and @xmath194 with the amplitude of the transition , that appears to be ` normal ' , namely @xmath188 . in the latter transition the released energy is only 332mev , and according to the discussed approach it should be absolutely dominated by the @xmath41 wave amplitude given by eq.([spp ] ) . also any fsi effects in this decay should be very small due to the proximity of the pions to the threshold . to certain extent this approach can be tested by comparing the rates of the transitions with the charged pions and with the neutral , by using respectively the mass of the charged and neutral pion in considering each of these decay modes . the effect of the pion mass difference is quite essential due to the small available energy , so that after numerical integration of the rate calculated from eq.([spp ] ) one finds , in place of the isotopic ratio 2 , the estimate 1.26 , [ yppprat ] which is in the agreement with the data @xcite : @xmath195 = ( 2.8 \pm 0.6)\%$ ] and @xmath196 = ( 2.00 \pm 0.32)\%$ ] . used is 0.146 . due to the small energy in the process , the estimate is quite insensitive to this particular value as long as @xmath139 is small . ] the integral over the spectrum of either of the modes of transition @xmath197 can then be compared with the result of the numerical integration over the spectrum of the decay @xmath19 produced by the amplitudes in the equations ( [ spp ] ) - ( [ d3 ] ) which gives the ratio of the rates of the two transitions from @xmath182 in terms of the ratio of the squares of the corresponding amplitudes @xmath53 and can be compared with the data . performing this calculation with the values of the parameters @xmath198 and @xmath199 used in the plots of fig.2 , gives the estimate 0.06 , _ m^ ( ) _ 0^ ( ) 0.04 , [ aest ] which certainly falls in the range of naturally expected magnitude of the relativistic terms in bottomonium : @xmath200 , and also quantifies the dynamical suppression of the leading nonrelativistic amplitude @xmath53 in the transition @xmath184 . however , given the large energy release in this transition it is rather difficult to quantify at present the fsi effect on the estimate ( [ aest ] ) . in either event , the suggested explanation of the observed peculiarity of the decay @xmath25 , heavily relies on the relative prominence of the spin - dependent @xmath44 waves , @xmath127 and @xmath128 , in this decay , as can be seen from the plots of fig.2 . the relative contribution of these two waves is the largest near the minimum in the invariant mass distribution at @xmath201gev , which is the main experimentally testable qualitative prediction of the suggested mechanism . it should be mentioned in connection with the suggested presence of the spin - dependent @xmath44 waves that an experimental search for quarkonium polarization effects in @xmath25 has been done @xcite . it was found that the data were consistent at 65% confidence level with the expectation that the daughter @xmath202 retains the polarization of the parent @xmath182 along the beam axis " . it thus appears that the test has not been sensitive enough and a more detailed study of the bottomonium polarization effects in this transition is needed . in summary . within the standard approach to two - pion transitions in heavy quarkonium , based on the multipole expansion in qcd and chiral dynamics of the pions , the first relativistic terms of order @xmath22 in the transition amplitudes arise from the effects of the @xmath203 mixing in the part determined by the e1 interaction and from the chromo - magnetic m1 interaction . these terms are parametrized by the quantities @xmath186 ( eq.([asd ] ) ) and @xmath187 ( eq.([amg ] ) ) . the significance of the chromo - magnetic term in two - pions transitions in charmonium can be approximately estimated from the available data on the dipion invariant mass spectrum in the decays @xmath161 and @xmath24 . the estimated value is 0.15 - 0.2 of the dominant nonrelativistic amplitude . the effect of the chromo - magnetic term is enhanced in the yet unobserved transition @xmath204 and is expected to significantly distort the spectrum in the latter decay and also result in a suppression of its rate , as shown in fig.1 . the absolute determination of the leading nonrelativistic amplitude in charmonium , the chromo - polarizability , is of interest for other applications , e.g. the charmonium scattering on nuclei . such determination generally suffers from fsi effects of the two pion rescattering . these effects are estimated as a next term in the chiral expansion and amount to only few percent at a low invariant mass of the two - pion system near the threshold . an extrapolation to higher values of @xmath153 can be done using experimental data . with the presently available data there is no indication of a large fsi effect . furthermore , an agreement of the chiral - limit formulas with the data @xcite on the @xmath44 wave in the decay @xmath205 , as well as the agreement with the data of the theoretical prediction @xcite for the ratio of the rates of the decays @xmath5 and @xmath206 , suggest that the fsi effects may amount to at most a moderate fraction of the amplitude of the transition in charmonium . an experimental measurement of the ratio of the @xmath127 wave ( eq.([d2 ] ) ) and the @xmath128 wave ( eq.([d3 ] ) ) in @xmath206 would provide a direct test of the fsi effect . the relativistic terms in the two - pion transitions may hold the clue to solving the puzzle of the unusual dipion mass spectrum observed in the transition @xmath25 , if the leading nonrelativistic quarkonium matrix element in this transition is strongly suppressed due to details of bottomonium wave functions . although a detailed quantitative description of the latter decay is not yet attainable within the present knowledge , the suggested mechanism necessarily predicts a noticeable presence of the polarization - dependent @xmath127 and @xmath128 waves in the amplitude , which prediction can be tested experimentally . i thank misha shifman and arkady vainshtein for illuminating discussions . this work is supported in part by the doe grant de - fg02 - 94er40823 . 99 g. s. abrams _ et . lett . * 34 * , 1181 ( 1975 ) . m. b. voloshin , jetp lett . * 21 * , 347 ( 1975 ) [ pisma zh . * 21 * , 733 ( 1975 ) ] . l. s. brown and r. n. cahn , phys . rev . lett . * 35 * , 1 ( 1975 ) . k. gottfried , phys . * 40 * , 598 ( 1978 ) . m. b. voloshin and v. i. zakharov m. b. voloshin , phys . b * 562 * , 68 ( 2003 ) [ arxiv : hep - ph/0302261 ] . a. sibirtsev and m. b. voloshin , phys . rev . d * 71 * , 076005 ( 2005 ) [ arxiv : hep - ph/0502068 ] . a. b. kaidalov and p. e. volkovitsky , phys . lett . * 69 * , 3155 ( 1992 ) . j. z. bai _ et al . _ [ bes collaboration ] , phys . d * 62 * , 032002 ( 2000 ) [ arxiv : hep - ex/9909038 ] . j. p. alexander _ et al . _ [ cleo collaboration ] , phys . d * 58 * , 052004 ( 1998 ) [ arxiv : hep - ex/9802024 ] . f. butler _ et al . _ [ cleo collaboration ] , phys . d * 49 * , 40 ( 1994 ) . s. e. csorna _ et al . _ [ cleo collaboration ] , report cleo conf 02 - 06 , arxiv : hep - ex/0207060 . f. k. guo , p. n. shen and h. c. chiang , arxiv : hep - ph/0604252 . m. b. voloshin , nucl . b * 154 * , 365 ( 1979 ) . m. b. voloshin , mod . a * 19 * , 665 ( 2004 ) [ arxiv : hep - ph/0402011 ] . j. l. rosner , phys . d * 64 * , 094002 ( 2001 ) [ arxiv : hep - ph/0105327 ] . g. belanger , t. a. degrand and p. moxhay , phys . d * 39 * , 257 ( 1989 ) . r. s. chivukula , a. g. cohen , h. georgi and a. v. manohar , phys . b * 222 * , 258 ( 1989 ) . r. s. chivukula , a. g. cohen , h. georgi , b. grinstein and a. v. manohar , annals phys . * 192 * ( 1989 ) 93 . h. albrecht _ et al . _ [ argus collaboration ] , z. phys . c * 35 * , 283 ( 1987 ) .	two - pion transitions in charmonium and bottomonium are considered with the leading relativistic effects taken into account . the contribution of the chromo- magnetic interaction of the charmed quarks to the amplitude of the decay @xmath0 is estimated from the available data . this contribution is enhanced by a factor of three in the decay @xmath1 and should produce a noticeable modification of the rate and the spectrum in the latter decay . it is argued that the peculiar observed spectrum in the decay @xmath2 arises due to a dynamical suppression of the leading nonrelativistic quarkonium amplitude and thus enhanced prominence of the relativistic terms . also discussed are the effects of the final state interaction between the pions . * william i. fine theoretical physics institute + university of minnesota + * tpi - minn-06/22-t + umn - th-2508 - 06 + june 2006 + * two - pion transitions in quarkonium revisited + * william i. fine theoretical physics institute , university of minnesota , + minneapolis , mn 55455 + and + institute of theoretical and experimental physics , moscow , 117218 +
many games with a mathematical flavor involve moving blocks or balls according to some simple rotational or translational rule in an attempt to put them into some specified pattern . generally not all permutations of the blocks are possible , and potential moves overlap at only a few elements . for instance , with the rubik s cube the arrangement of the center squares on the faces is constrained and two rotations can affect at most @xmath11 pieces in common , while in hungarian ring puzzles only @xmath12 of the marbles are shared by both rotations . the @xmath13puzzle involves very small moves : tiles numbered @xmath14 through @xmath13 are placed in a @xmath15 grid , and the only moves involve sliding a tile into the empty spot , so that only two tiles are affected by each move . given that in each case small moves and limited overlap are intended primarily to produce games that are more easily playable by humans , it is natural to ask what would happen in the opposite extreme . in particular , would a rotational block - style puzzle with only a few , highly overlapping , rotations result in a large or a small number of accessible permutations of the blocks ? an extreme case of this is to consider rotating @xmath16 blocks of square tiles in a @xmath17 rectangle , or an overlapping union of such rectangles . for instance , figure [ fig : floor ] illustrates a tile arrangement of three @xmath17 rectangles for @xmath18 , along with three potential rotations of @xmath19 blocks of tiles . ( there are three other potential rotations in this particular arrangement . ) we find that in general somewhat surprisingly nearly all permutations of the tiles are possible , despite the minimal overlap shared by the @xmath17 regions . block rotations generate all @xmath20 permutations of blocks.,height=192 ] we use the notation @xmath21 to denote the group generated by all the possible rotations of @xmath0 squares in the tile arrangement , where the @xmath22 represent the generators , and refer to this as the _ puzzle group_. the pattern of tiles is _ admissible _ if it can be formed by overlapping @xmath1 rectangles and/or @xmath23 rectangles in a sequence such that if @xmath24 is even each new rectangle overlaps with the previous arrangement by at least one tile , and if @xmath24 is odd then it overlaps by at least two adjacent tiles . when there are @xmath2 total tiles , the @xmath0 rotations generate a subgroup of @xmath6 . furthermore , when @xmath24 is odd there are at least two disjoint orbits : with the tile arrangement colored like a checkerboard , the puzzle tiles permute like - colored squares . ( see figure [ fig : orbits ] . ) is odd.,height=144 ] additional restrictions on @xmath25 are immediately apparent . first , when @xmath26 , then each rotation results in an even permutation ( of @xmath27 four - cycles ) , and so @xmath28 . second , when @xmath24 is odd , then each rotation results in two orbits of @xmath29 four - cycles each . so assuming the entire union of @xmath1 rectangles is colored as a checkerboard , with @xmath3 squares shaded and @xmath4 squares unshaded , then if @xmath30 , then @xmath31 , while if @xmath32 , then @xmath33 , that is , @xmath25 is a subset of the even permutations in @xmath34 . in this paper we prove that in fact these are the only restrictions on the feasible permutations , except in two small cases . [ thm : main ] if @xmath35 then the puzzle group of an admissible figure on @xmath2 blocks is given by : 1 . if @xmath24 is even and @xmath36 then @xmath37 2 . if @xmath24 is odd and @xmath38 then shade the figure as a checkerboard in black and white . if there are @xmath3 elements in black and @xmath4 in white then @xmath39 where @xmath40 is the set of all even permutations in @xmath34 . 3 . if @xmath41 , then @xmath42 for the projective linear group under an appropriate labeling of vertices . + if @xmath10 and @xmath43 , then @xmath44 where the projection of @xmath25 onto each orbit is @xmath45 our proof is modeled after wilson s approach to finding the permutation group for a generalization of the @xmath13puzzle problem @xcite . in order to explain the method we require a few definitions . recall that a permutation group @xmath25 acting on set @xmath46 is _ transitive _ if it can send any @xmath47 to any @xmath48 ( i.e. @xmath49 ) , while it is _ primitive _ if it is transitive and does not preserve any bipartition ( i.e. @xmath50 ) . in particular , a _ doubly transitive _ group ( transitive with @xmath51 transitive for some @xmath47 ) is primitive . _ jordan s theorem _ says that a primitive group @xmath25 containing a @xmath11-cycle is either @xmath5 or @xmath6 ( e.g. theorem 13.3 of @xcite ) . with this in mind , our approach to proving the theorem is to first show that @xmath25 is doubly transitive on each of its orbits , and then show that @xmath25 contains a 3-cycle on each orbit ( with two exceptions ) . it follows that @xmath25 contains the product of the alternating groups on the orbits , which leaves only a small number of potential groups to consider . we first consider the most basic type of puzzle group , that on a @xmath52 rectangle . the general case will be derived from this at the end of the proof . there are only two generators to consider ; denote the generator on the left by @xmath53 , i.e. a clockwise rotation of the left @xmath16 square region , and the one on the right by @xmath54 , i.e. a clockwise rotation of the right @xmath16 square region . the puzzle group is @xmath55 , the group generated by @xmath53 and @xmath54 . label the tiles of the rectangle by their cartesian coordinates @xmath56 , with the tile in the upper left corner denoted @xmath57 , the tile to its right denoted @xmath58 , and so forth . then @xmath59 , for instance . let @xmath60 denote conjugation , so that @xmath61 denotes the location of tile @xmath62 after @xmath63 . the conjugate is relatively easy to compute by using the property that if @xmath64 then @xmath65 . it will be useful to note the actions of the generators : @xmath66 the most common approach to proving puzzle groups is to work with a _ commutator _ : @xmath67=g\,h\,g^{-1}\,h^{-1}\ ] ] this tends to involve simple shifts from the identity that can be easier to work with than the original action in the puzzle . we use a few commutators when showing double - transitivity . @xmath68(i , j ) & = & ( i , j-2 ) \textrm{\qquad when \qquad $ i>1$\ \ and\ \ $ j>2 $ } \\ { [ } \sigma_\l,\sigma_\r{]}(i , j ) & = & ( i+2,j ) \textrm{\qquad when \qquad $ i < k-1$\ \ and\ \ $ 1<j < k+1$}\end{aligned}\ ] ] we are not concerned with the action outside the specified regions , so we do not describe it here . for construction of @xmath11cycles we find that other expressions which take into account the order of the rotations can be easier to work with . in this section we develop those building blocks . the two simplest formulas we can derive are simply rotations by @xmath69 , so @xmath70 and @xmath71 . more generally , since @xmath53 and @xmath54 send a tile @xmath72 to nearly the same location , then a product of four @xmath53 and @xmath54 terms will involve only minor shifts for most tiles . we write out a handful of such expressions and then combine them to get @xmath11cycles . a simple example we will work with is @xmath73 tiles are shifted to the right by @xmath74 , and when this wraps around the boundary then they are also flipped vertically . another useful case is @xmath75 most tiles are shifted down and to the right by one diagonally . the action of @xmath76 is a bit more complicated : @xmath77 the location of @xmath72 in the table is the location it is mapped to . this shows that a cycle formed by the left side , top , and right side of the rectangle rotates clockwise by @xmath78 tiles , and the rest remains fixed . [ lem : double - transitive ] the puzzle group for a @xmath17 rectangle is doubly transitive on each orbit . we begin with the case when @xmath24 is odd . let @xmath79 be the set of tiles with @xmath80 even ( the shaded region in figures [ fig : orbits ] and [ fig : trans - induction ] ) . this is one of the two orbits of @xmath25 on the set of tiles . since @xmath81 is just the reflection of @xmath79 through the centerpoint @xmath82 and @xmath83 then double - transitivity of @xmath79 also implies double - transitivity on the set of tiles with @xmath80 odd . let @xmath84 be the square immediately to the left of the center of @xmath54 ; this is the center of the @xmath53 rotation . rotations preserve the parity of blocks , i.e. if @xmath85 then @xmath86 and @xmath87 , so @xmath88 . given @xmath89 in @xmath90 let @xmath91 and for @xmath92 define @xmath93 this is just those tiles with both @xmath94 and @xmath95 coordinates at most @xmath89 from the center of the @xmath54 rotation . we show that @xmath96,\ [ \sigma_\l,\sigma_\r]\rangle\ , s_{d-1 } \supseteq s_d$ ] , and so by induction @xmath96,\ [ \sigma_\l,\sigma_\r]\rangle\,a \supseteq e\setminus \{(1,1)\}$ ] . since @xmath96,\ [ \sigma_\l,\sigma_\r]\rangle \leq stab(1,1)$ ] then @xmath97 . for the base case , when @xmath98 then @xmath99 and @xmath100 , so the claim is trivial . [ fig : boundaries ] \,s_2 $ ] contains the left boundary of @xmath101.,height=168 ] for the inductive step assume that @xmath102 . if @xmath103 and @xmath104 then @xmath105(i , j)=(i , j-2)$ ] , and so @xmath105\,s_{d-1}$ ] includes the left boundary of @xmath106 . rotation of the left boundary under @xmath107 includes all of @xmath108 . the final case , when @xmath109 , does not require the rotation and completes the proof that @xmath97 . when @xmath24 is even let @xmath79 be the set of all tiles and @xmath110 . then @xmath111 trivially . the method of proof is the same , but in the inductive step @xmath105\,s_{d-1}$ ] now misses the top and bottom of the left boundary of @xmath106 when @xmath112 . however , @xmath105\,[\sigma_\l,\sigma_\r](i , j)=(i+2,j-2)$ ] when @xmath113 and @xmath114 , and so if @xmath112 then @xmath96,\ [ \sigma_\l,\sigma_\r]\rangle\ , s_{d-1}$ ] also contains the lower left tile of @xmath106 . applying powers of @xmath54 to this covers @xmath108 . having established double transitivity , we now seek a 3cycle . [ lem:3-cycle ] if @xmath115 then there is a 3cycle in the @xmath17 rectangle with generators @xmath53 and @xmath54 . more precisely : 1 . if @xmath24 is even then @xmath116 is a @xmath11cycle where @xmath117 when @xmath26 and @xmath118 when @xmath119 . if @xmath24 is odd then @xmath120 is a @xmath11cycle where @xmath121 and @xmath122 is the order of @xmath123 . a 3cycle on the other orbit may be obtained by swapping @xmath54 and @xmath53 . * case 1 ( @xmath24 even ) : * we consider @xmath26 . the methodology when @xmath119 is the same , but with different cycle structures . the actions of @xmath124 and @xmath125 were described earlier . the exponent in the conjugated term is @xmath126 it is a short exercise to verify that @xmath127 consists of * @xmath14-cycles : @xmath128 where @xmath129 * @xmath130-cycles : @xmath131 where @xmath132 the cycle structure of @xmath133 then consists of * @xmath14cycles : @xmath128 where @xmath134 * @xmath11cycle : @xmath135 * @xmath130cycles : one @xmath130 cycle for each tile @xmath136 * @xmath137cycle : one cycle containing @xmath138 and @xmath139 , among others . this gives @xmath140 tiles , and so it is the complete cycle structure . taking the @xmath141 power leaves only the square of the @xmath11cycle , which is also a @xmath11cycle . * case 2 ( @xmath24 odd ) :* it suffices to show existence of a @xmath11cycle on the orbit @xmath142 since reflecting the @xmath52 region through its centerpoint , i.e. @xmath82 and @xmath83 , swaps @xmath79 and @xmath81 as well as @xmath53 and @xmath54 , transforming the @xmath11cycle on @xmath79 into a @xmath11cycle on @xmath81 . the main term is @xmath143 and so @xmath144 our theorem uses exponents @xmath145 . when @xmath146 then acts as : @xmath147 when @xmath148 then an inductive argument shows that acts as : @xmath149 * main case : @xmath150 * set @xmath151 , in accordance with lemma [ lem:3-cycle ] . there is a 3cycle @xmath152 , a @xmath137cycle containing @xmath57 and @xmath153 other tiles , an @xmath154cycle containing @xmath155 and @xmath156 other tiles , two @xmath130cycles with one containing @xmath157 and the other @xmath158 , and a fixed tile @xmath159 . we now specialize further : * subcase of @xmath160 : * since @xmath151 then the most common transition is when @xmath161 and @xmath162 , in which case @xmath163 starting at @xmath164 , where @xmath165 , this induces a long sequence alternating between two types of @xmath95 terms : @xmath166 the pattern does nt hold if @xmath167 or @xmath168 , which is why we do nt allow either . in total the sequence contains @xmath169 terms , ending in @xmath170 . after this comes a short three - term sequence @xmath171 . the tile @xmath172 satisfies the requirements for once again , leading to another sequence using pattern , this time with @xmath173 additional terms , ending in @xmath174 . following this is another three - term sequence , @xmath175 . once again repeatedly apply to get @xmath176 additional terms , ending with @xmath177 . finally , following this is another three - term sequence , @xmath178 . but @xmath164 is just what we started with , and so we are done . the total number of terms in the cycle is then @xmath179 . there is only one term of the form @xmath164 in each sequence , so every @xmath164 makes a distinct such sequence , and in particular there are @xmath180 such cycles of order @xmath181 . this , along with the @xmath182 cycles listed before our restriction to @xmath160 , accounts for all @xmath140 tiles . since @xmath183 then @xmath184 is not divisible by @xmath11 , and taking the @xmath185 power of then leaves only a @xmath11cycle . * subcase of @xmath186 : * this is nearly identical to the proof when @xmath160 , but with sequences ending at slightly different values . however , again there are @xmath180 cycles of order @xmath181 , each containing a member of @xmath187 , and so again there is only one cycle of order divisible by @xmath11 . * subcase of @xmath188 : * starting at a tile of the form @xmath189 , equation gives a sequence of @xmath190 subsequent terms , ending on @xmath191 . this is followed by @xmath192 , for a total of @xmath193 terms in each such cycle . this gives a family of @xmath194 cycles of order @xmath193 which , when combined with the cycles given before case 2.1 , accounts for all @xmath140 tiles . since @xmath195 then @xmath196 is not divisible by @xmath11 , and so once again there is only one cycle of order divisible by @xmath11 . * secondary case : @xmath197 * the approach used when @xmath150 still applies . however , some of the cycles previously found have order divisible by @xmath11 when @xmath197 , so a different @xmath198 will be needed . in fact , numerous subcases with different exponents @xmath198 are required in order to avoid cycles of order divisible by @xmath11 , and these subcases tend to have many more cycle types . following is a chart explaining the cycle structure for these remaining cases . in each case the cycle type is listed along with exactly one tile from each such cycle . the simplest case is when @xmath199 . @xmath200 when @xmath201 then the theorem uses @xmath146 . the group action in this case is given by . going through the cycle structure , as in the @xmath151 case , we find it to be : @xmath202 when @xmath203 there is a common set of cycle types , plus additional cycles depending on the value of @xmath24 modulo @xmath204 . @xmath205 @xmath206 recall that our proof will utilize the fact that a subgroup of @xmath6 which contains a @xmath11cycle and is doubly transitive is either @xmath5 or @xmath6 . the first step is to show that for any admissible figure , @xmath25 is doubly transitive on its orbit(s ) . by lemma [ lem : double - transitive ] @xmath25 is doubly transitive on its orbits for a single @xmath207 or @xmath23 rectangle . proceeding by induction , assume that @xmath25 is doubly transitive on all of its orbits for an admissible figure constructed from @xmath208 rectangles of dimension @xmath207 or @xmath23 . add another @xmath207 or @xmath23 rectangle to this figure so that the resulting figure , which is constructed from @xmath209 such rectangles , is admissible . there are at least @xmath24 tiles which belong to the original figure but not to the added rectangle . choose one of these tiles and call it @xmath47 . by hypothesis , the stabilizer of @xmath47 in the original figure is transitive , while the generators of the new rectangle are transitive . since the original figure and the new rectangle overlap in the orbit of @xmath47 , but not at @xmath47 itself , then the stabilizer of @xmath47 in the new figure is also transitive and the figure is doubly transitive on the orbit of @xmath47 . likewise , if @xmath24 is odd then @xmath47 can be chosen to be in either of the two orbits , so the figure is doubly transitive on each orbit . hence for any admissible figure , @xmath25 is doubly transitive on its orbits . next we combine double - transitivity and the @xmath11cycles already proven to exist . this greatly limits the number of groups that are possible , and we then refine this down to a single possible answer in each case . * case 1 ( @xmath115 is even ) :* lemma [ lem:3-cycle ] shows that @xmath25 contains a 3cycle . the previous paragraph establishes that @xmath25 is doubly transitive , and so @xmath210 or @xmath6 . since @xmath24 is even then each generator consists of @xmath211 disjoint @xmath130cycles . as a result , if @xmath26 then the generators are even permutations , and so @xmath28 , implying @xmath210 . if @xmath119 then the generators are odd permutations and so @xmath212 , implying @xmath213 . * case 2 ( @xmath115 is odd ) :* there are two orbits , of some @xmath3 and @xmath214 tiles each , and so @xmath215 . from lemma [ lem:3-cycle ] there is a @xmath11cycle @xmath216 on the @xmath3-element orbit . the proof of jordan s theorem generates @xmath217 by conjugating this specific @xmath11cycle and multiplying the resulting terms . since @xmath218 then jordan s theorem implies that @xmath219 . likewise , @xmath220 . hence @xmath221 . each generator consists of @xmath222 disjoint @xmath130cycles , exactly @xmath223 in each of the two orbits . if @xmath30 then @xmath223 is even , and so the @xmath224 disjoint @xmath130cycles in each orbit make an even permutation in that orbit . it follows that @xmath31 , and so in fact @xmath225 . if @xmath32 then @xmath223 is odd and so the generators @xmath53 and @xmath54 act on each orbit as an odd permutation , but are themselves even permutations , and so @xmath226 . however , the only group satisfying @xmath226 is @xmath227 . to see this observe that since @xmath33 then every @xmath228 acts as an even permutation on both orbits or as an odd permutation on both orbits . it follows that if @xmath229 and @xmath230 then they act as odd permutations on both orbits , and therefore @xmath231 acts as an even permutation on each orbit , i.e. @xmath232 and so @xmath233 . * case 3 ( @xmath234 ) :* when @xmath235 ( a @xmath236 rectangle ) then the generators are @xmath237 equivalently , if we label @xmath58 as @xmath238 and then number from @xmath239 to @xmath130 counterclockwise starting at @xmath240 and ending at @xmath241 , then the generators are @xmath242 and @xmath243 . an alternate set of generators is @xmath244 and @xmath245 . the projective group @xmath246 includes the following transformations on @xmath247 : @xmath248 the generator @xmath249 is the transformation @xmath250 , while @xmath251 is the transformation @xmath252 , so @xmath253 . those two transformations in fact generate @xmath246 ( e.g. @xcite ) , and so under an appropriate labeling of vertices then @xmath254 . suppose instead that @xmath255 . the proof of case 1 carries through as long as there is a @xmath11cycle . the construction of every admissible figure with @xmath255 starts by overlapping two regions of sizes @xmath236 and/or @xmath256 . up to symmetry ( rotation or reflection through an axis ) this region will contain either a @xmath257 , a @xmath236 joined at a corner square to a @xmath258 , or two @xmath236 joined at a @xmath259 angle to create a @xmath260 missing a corner . more concretely , let @xmath261 and @xmath262 be the left and right generators defined in . a third generator @xmath263 and a @xmath11cycle will now be designated in each of the three cases just discussed : @xmath264\right)^2 = ( ( 1,2),\ , ( 1,3),\ , ( 2,4 ) ) \\ \sigma_3 & = & ( ( 2,3),\ , ( 2,4),\ , ( 3,4),\ , ( 3,3 ) ) \\ \textrm{3 - -cycle } & & [ \sigma_3,\sigma_2 ] = ( ( 2,2),\,(2,3),\,(2,4 ) ) \\ \sigma_3 & = & ( ( 2,2),\ , ( 2,3),\ , ( 3,3),\ , ( 3,2 ) ) \\ \textrm{3 - -cycle } & & [ \sigma_3,\sigma_1 ] = ( ( 2,1),\ , ( 2,2),\ , ( 2,3))\end{aligned}\ ] ] * case 4 ( @xmath265 ) :* when @xmath266 then the pair of generators are @xmath267 consider action on the orbit @xmath268 . the puzzle group @xmath25 is doubly - transitive on each orbit and contains the @xmath11cycle @xmath269 , and so @xmath270 . but @xmath271 contains the odd permutation @xmath272 , and so @xmath273 . it can be verified by brute force ( e.g. gap or mathematica or a very long exercise ) that @xmath274 , and so in fact @xmath275 . when @xmath276 then once again start with a @xmath277 region and attach a @xmath277 or @xmath278 to make a larger admissible figure . this time two @xmath11cycles are needed , one in the orbit @xmath79 and another in the orbit @xmath81 . up to symmetry ( rotation or reflection through an axis ) this figure will contain either a @xmath279 , a @xmath277 joined by two squares near a corner to a @xmath260 ( two cases ) , or two @xmath277 joined at a @xmath259 angle to create a @xmath15 missing a corner . more concretely let @xmath261 and @xmath262 be the left and right generators defined in . a third generator @xmath263 and a @xmath11cycle on each orbit will now be designated in each of the four cases just mentioned . @xmath280 ; % comm[g1,g3 ] % permutationproduct[g1,g2,g2,g2,g3,g3 ] \sigma_3 & = & ( ( 1,3),\ , ( 1,5),\ , ( 3,5),\ , ( 3,3))\,((1,4),\ , ( 2,5),\ , ( 3,4),\ , ( 2,3 ) ) \\ & & \textrm{3 - -cycles } \ [ \sigma_1,\sigma_3]^2 = ( ( 1,4),\,(3,2),\,(2,3 ) ) \\ & & \ \textrm { and } \ ( \sigma_3 ^ 2\sigma_2^{-1}\sigma_1)^{20 } = ( ( 1,3),\ , ( 3,5),\ , ( 2,4 ) ) \\ % % mathematica % g3 : = cycles[{{8 , 14 , 19 , 17 } , { 13 , 16 , 18 , 12 } } ] ; % permutationproduct[f[g1,g2],g3,g3 ] % permutationproduct[f[g2 , g3],g1,g1 ] \sigma_3 & = & ( ( 2,4),\,(2,6),\,(4,6),\,(4,4))\,((2,5),\,(3,6),\,(4,5),\,(3,4 ) ) \\ & & \textrm{3 - -cycles } \ ( \sigma_3 ^ 2\,[\sigma_2 ^ 2,\sigma_1 ^ 2])^2 \ \textrm { and } \ ( \sigma_1 ^ 2\,[\sigma_3 ^ 2,\sigma_2 ^ 2])^4 \\ % % mathematica % g3 : = cycles[{{11 , 13 , 19 , 17 } , { 12 , 16 , 18 , 14 } } ] ; % comm[g1 , g3 ] % permutationproduct[comm[g2 , g3 ] , comm[g1 , g3 ] ] \sigma_3 & = & ( ( 3,3),\,(3,5),\,(5,5),\,(5,3))\,((3,4),\,(4,5),\,(5,4),\,(4,3 ) ) \\ & & \textrm{3 - -cycles } \ [ \sigma_1,\sigma_3]^4 \ \textrm { and } \ \left([\sigma_1,\sigma_3]\ , [ \sigma_2,\sigma_3 ] \right)^2 \\ % % % mathematica - this is just an l shape , which is covered in the ( 2,2 ) case . % % g3 : = cycles[{{9 , 11 , 18 , 16 } , { 10 , 15 , 17 , 13 } } ] ; % % comm[g1 , g3 ] % % permutationproduct[comm[g2 , g3 ] , comm[g2 , g3 ] , comm[g1 , g3 ] , comm[g1 , g3 ] ] % \sigma_3 & = & ( ( 3,1),\,(3,3),\,(5,3),\,(5,1))\,((3,2),\,(4,3),\,(5,2),\,(4,1 ) ) \\ % & & \textrm{3 - -cycles } \ [ \sigma_1,\sigma_3]^2 % \ \textrm { and } \ \left([\sigma_1,\sigma_3]^2\,[\sigma_2,\sigma_3]^2 \right)^2 % % % mathematica % g3 : = cycles[{{6 , 8 , 15 , 13 } , { 7 , 12 , 14 , 10 } } ] ; % permutationproduct[f[g1,g3],g2 ] % permutationproduct[g1 , g1 , g2 , g2 , g3 ] \sigma_3 & = & ( ( 2,2),\ , ( 2,4),\ , ( 4,4),\ , ( 4,2))\,((2,3),\ , ( 3,4),\ , ( 4,3),\ , ( 3,2 ) ) \\ & & \textrm{3 - -cycles } \ ( \sigma_2\,[\sigma_3 ^ 2,\sigma_1 ^ 2])^4 \ \textrm { and } \ ( \sigma_3\sigma_2 ^ 2\sigma_1 ^ 2)^{20 } % % mathematica - could n't get this to work , but ( 3,1 ) rotation and ( 2,2 ) rotation cover this case . % % g3 : = cycles[{{5 , 7 , 15 , 13 } , { 6 , 11 , 14 , 9 } } ] ; % % permutationproduct[g1 , f[g2 , g3 ] ] % % ? ? % \sigma_3 & = & ( ( 2,1),\,(2,3),\,(4,3),\,(4,1))\,((2,2),\,(3,3),\,(4,2),\,(3,1 ) ) \\ % & & \textrm{3 - -cycles } \ ( [ \sigma_3 ^ 2,\sigma_2 ^ 2]\sigma_1)^4 % \ \textrm { and } \ ? % % mathematica - unnecessary , the other rotation is already a case . % % g3 : = cycles[{{7 , 13 , 17 , 15 } , { 8 , 14 , 16 , 11 } } ] ; % % permutationproduct[comm[g1,g2],g3 ] % % permutationproduct[f[g1,g3],g2,g2 ] % \sigma_3 & = & ( ( 2,3),\,(2,5),\,(4,5),\,(4,3))\,((2,4),\,(3,5),\,(4,4),\,(3,3 ) ) \\ % & & \textrm{3 - -cycles } \ ( \sigma_3\,[\sigma_2,\sigma_1])^8 % \ \textrm { and } \ ( \sigma_2 ^ 2[\sigma_3 ^ 2,\sigma_1 ^ 2])^4 % \\ \end{aligned}\ ] ] in some cases the generator @xmath263 may not appear in either of the regions being overlapped , such as when overlapping two @xmath277 regions to make a @xmath281 region . however , we are studying the group generated by all the possible rotations of @xmath0 squares in the tile arrangement , and so @xmath263 is still a valid rotation in the union of the two @xmath277 regions . we are grateful to the mathematics department at texas christian university for hosting the reu during which this work was done . special thanks to george t. gilbert and rhonda l. hatcher for their guidance and to david addis for his assistance with gap . thanks also go to craig morgenstern who patiently fixed the disasters we caused to his computer systems . we particularly appreciate the efforts of an anonymous referee .	this paper discusses the permutations that are generated by rotating @xmath0 blocks of squares in a union of overlapping @xmath1 rectangles . it is found that the single - rotation parity constraints effectively determine the group of accessible permutations . if there are @xmath2 squares , and the space is partitioned as a checkerboard with @xmath3 squares shaded and @xmath4 squares unshaded , then the four possible cases are @xmath5 , @xmath6 , @xmath7 , and the subgroup of all even permutations in @xmath8 , with exceptions when @xmath9 and @xmath10 .
theoretical investigation of ruderman@xcite leads to the conclusion that matter may be anisotropic in densities of order @xmath0 . the impact of anisotropy on steller configuration may be found in pioneering works of bower and liang@xcite and herrera and santos@xcite . anisotropy may occur due to existance of type 3a superfluids@xcite , or phase transition@xcite . consenza _ _ et . al.__@xcite developed the procedure to obtain anisotropic solutions from isotropic solutions of einstein s field equations . tikekar and thomas@xcite found exact solutions of einstein s field equations for anisotropic fluid sphere on pseudo spheroidal spacetime . the key feature of their model is the high variation of density from centre to boundary of stellar configuration . the class of exact anisotropic solutions on spherically symmetric spacetime has been obtained by mak and harko@xcite . karmakar _ _ et . al.__@xcite analysed the role of pressure anisotropy for vaidya - tikekar@xcite model . paul _ _ et . al__@xcite developed anisotropic stellar model for strage star . a core - envelope model describing superdense stars with anisotropic fluid distribution has been obtained by thomas and ratanpal@xcite , thomas _ _ et . al.__@xcite & tikekar and thomas@xcite . hence the study of anisotropic fluid distribution is important in general theory of relativity . + + the study of einstein - maxwell system is carried out by several authors . patel and kopper@xcite obtained charged analog of vaidya - tikekar@xcite solution . the study of analytic models of quark stars have been carrried out by komathiraj and maharaj@xcite , they found a class of solutions of einstein - maxwell system . charged anisotropic matter with linear equation of state have been extensively studied by thirukkanesh and maharaj@xcite . + + hence , both anisotropy and electromagnatic field are important in relativistic astrophysics . in this paper charged anisotropic model of a stellar configuration have been studied on the background of paraboloidal spacetime . sec-2 describes , the field equations for charged static stallar configuration on paraboloidal spacetime and their solution is obtained . sec-3 described the physical plausibility condition and sec-4 contains discussion . the interior of stellar configuration is described by static spherically symmetric paraboloidal spacetime metric , @xmath1 with the energy - momentum tensor for anisotropic charged fluid , @xmath2 where @xmath3 is the energy density , @xmath4 is the radial pressure , @xmath5 is the tangential pressure and @xmath6 is the electric field intensity . these quantities are measured relative to the comoving fluid velocity @xmath7 . for the spacetime metric ( [ ispacetime1 ] ) and energy - momentum tensor ( [ emtensor ] ) , the einstein - maxwell system takes the form , @xmath8 @xmath9 @xmath10-\frac{\nu ' r}{2r^{2}\left(1+\frac{r^{2}}{r^{2}}\right)^{2}}-\frac{1}{r^{2}\left(1+\frac{r^{2}}{r^{2}}\right)^{2}},\ ] ] @xmath11 where @xmath12 is the proper charge density and prime denotes differentiation with respect to @xmath13 . in field equations ( [ rho1 ] ) - ( [ sigma1 ] ) , velocity of light @xmath14 is taken as @xmath15 also @xmath16 . + + the anisotropic parameter @xmath17 is defined as , @xmath18 to solve the system ( [ rho1 ] ) - ( [ sigma1 ] ) radial pressure is assumed to be of the form , @xmath19 where @xmath20 is the model parameter and @xmath21 is central pressure . at the boundary of the star @xmath22 , @xmath4 must vanish , which gives @xmath22 as the radius of the star . + + this form of radial pressure is prescribed by sharma and ratanpal@xcite to describe anisotropic stellar model admitting a quadratic equation of state on paraboloidal spacetime . equations ( [ pr2 ] ) and ( [ pr1 ] ) gives , @xmath23 we assume electric field intensity of the form , @xmath24 where @xmath25 is a model parameter , from equation ( [ e1 ] ) it is clear that @xmath6 is decreasing in radially outward direction . equations ( [ nudash1 ] ) and ( [ e1 ] ) leads to , @xmath26 and hence , @xmath27+\left(\frac{1-p_{0}-k}{2}\right)\frac{r^{2}}{r^{2}},\ ] ] where @xmath28 is constant of integration . therefore spacetime metric ( [ ispacetime1 ] ) is written as , @xmath29 the spacetime metric ( [ ispacetime2 ] ) should continously match with reissner - nordstrom spacetime metric @xmath30 at the boundary of the star @xmath22 , where @xmath31 . this matching conditions gives @xmath32 and @xmath33 where @xmath34 is the mass enclosed the spherical body of radius @xmath35 , hence the electric field intensity parameter @xmath36 directly effects the mass of the star . equations ( [ pr1 ] ) , ( [ pt1 ] ) , ( [ delta1 ] ) and ( [ nudash2 ] ) gives anisotropic parameter @xmath17 as , @xmath37}{4r^{2}\left(1+\frac{r^{2}}{r^{2}}\right)^{3}},\ ] ] which vanishes at @xmath38 , where @xmath39 , @xmath40 , @xmath41 . equations ( [ rho1 ] ) and ( [ e1 ] ) gives , @xmath42 and from equations ( [ delta1 ] ) , ( [ pr2 ] ) and ( [ delta2 ] ) we get expression of @xmath5 as , @xmath43 hence equations ( [ rho2 ] ) , ( [ pr2 ] ) , ( [ pt2 ] ) , ( [ e1 ] ) and ( [ delta2 ] ) describes matter density , radial pressure , tangential pressure , electric field intensity and measure of anisotropy respectively . following delgaty and lake@xcite we impose the following condition on the sytem to make model physically acceptable . * @xmath44 for @xmath45 . * @xmath46 for @xmath45 . * @xmath47 for @xmath48 . * @xmath49 ; @xmath50 , for @xmath48 . from equation ( [ rho2 ] ) , @xmath51 and @xmath52 , therefore , @xmath53 for @xmath54 if @xmath55 , i.e. @xmath56 from equation ( [ pr2 ] ) @xmath57 as @xmath20 and @xmath58 . hence @xmath59 for @xmath54 . it is required that @xmath60 for @xmath54 and further to get the simple bounds on @xmath61 and @xmath36 , we assume @xmath62 at @xmath22 . + + from ( [ pt2 ] ) , @xmath63 as @xmath20 , and @xmath64 , at @xmath22 , @xmath65 if , @xmath66 but @xmath36 should be chosen such that @xmath61 is positive , which restrict the value of @xmath36 as , @xmath67 hence , @xmath68 which is condition for positivity of @xmath5 . + + hence condition ( i ) is satisfied throughout the star . for the values of @xmath36 and @xmath61 specified in ( [ k3 ] ) , programatically it has been varified that condition ( ii ) i.e. energy condition is satisfied throughout the star . from equation ( [ rho2 ] ) , @xmath69}{\left(1+\frac{r^{2}}{r^{2 } } \right)^{3}},\ ] ] from equation ( [ drhodr ] ) , @xmath70 and @xmath71 . hence @xmath3 is decreasing throught the star . from equation ( [ pr2 ] ) , @xmath72 now , @xmath73 and @xmath74 as @xmath20 . hence @xmath4 is decreasing throughout the star . from equation ( [ pt2 ] ) , @xmath75}{2r^{4}\left(1+\frac{r^{2}}{r^{2}}\right)^{4}},\ ] ] where @xmath76 , @xmath77 and @xmath78 . now @xmath79 and @xmath80 . substituting @xmath61 from equation ( [ p0 ] ) in @xmath81 , @xmath82 if @xmath83 . this further restrict value of @xmath36 as @xmath84 . + + hence , if @xmath85 then @xmath86 , @xmath87 and @xmath88 are decreasing in radially outward direction between @xmath54 . from equations ( [ drhodr ] ) and ( [ dprdr ] ) we have , @xmath89}.\ ] ] at the centre of the star @xmath90 if @xmath91 , which is consistent with condition ( [ k4 ] ) and at the boundary of the star @xmath92 if @xmath93 , which is also consistent with conditon ( [ k4 ] ) . from equations ( [ drhodr ] ) and ( [ dptdr ] ) we have , @xmath94}{4\left(1+\frac{r^{2}}{r^{2}}\right)\left[(5+k)+(1-k)\frac{r^{2}}{r^{2}}\right]}.\ ] ] now , @xmath95 if @xmath96 , which is consistent with condition ( [ k4 ] ) and @xmath97 if @xmath98 , which is also consistent with condition ( [ k4 ] ) . hence for @xmath99 and @xmath100 , all the physical plausibility conditions are satisfied . certain ascepts of charged relativistic star on paraboloidal spacetime have been discussed . it is observed that all the physical plausibility conditions are satisfied for @xmath101 and @xmath100 . the plots of @xmath3 ( charged , uncharged ) , @xmath4 & @xmath5 ( charged ) , @xmath4 & @xmath5 ( uncharged ) , anisotropy @xmath17 ( charged , uncharged ) , @xmath102 ( charged , uncharged ) , @xmath103 & @xmath104 ( charged ) , @xmath103 & @xmath104 ( uncharged ) over @xmath105 for @xmath106 , @xmath107 and taking @xmath108 are shown in figures 1 to 7 . it is observed that energy condition is satisfied throughout the star . when @xmath109 the value of @xmath110 and the model reduces to the sharma & ratanpal@xcite model . hence the model described here is charged generalization of particular case @xmath110 of uncharged sharma & ratanpal@xcite model . r. ruderman , _ annu . astrophys . _ * 10 * ( 1972 ) 427 . r l bowers and e p t liang , _ astrophysical j. _ * 188 * ( 1974 ) 657 . l herrera and n o santos , _ phys . * 286 * ( 1997 ) 53 . r kippenhahn and a weigert , stellar structure and evolution ( springer verlang , berlin heidelberg , new york ) ( 1990 ) . a i sokolov , _ jetp _ * 52 * ( 1980 ) 575 . m cosenza , l herrera , m esculpi and l witten , _ j. math . * 22 * ( 1981 ) 118 . r tikekar and v o thomas , _ pramana j. phys . _ * 52 * ( 1999 ) 237 . m k mak and t harko , _ proc . a _ * 459 * ( 2003 ) 393 . s karmakar , s mukherjee , r sharma and s d maharaj , _ pramana j. phys . _ * 68 * ( 2007 ) 881 . p c vaidya and r tikekar , _ j. astrophys . astr . _ * 3 * ( 1982 ) 325 . b c paul , p k chattopadhyay , s karmakar and r tikekar , _ mod . * 26 * ( 2011 ) 575 . v o thomas and b s ratanpal , _ int . * 16 * ( 2007 ) 1479 . v o thomas , b s ratanpal and p c vinodkumar , _ int . j. mod * 14 * ( 2005 ) 85 . r tikekar and v o thomas , _ pramana j. phys . _ * 64 * ( 2005 ) 5 . l k patel and s s kopppar , _ ast . _ * 40 * ( 1987 ) 441 . k komathiraj and s d maharaj , _ int . j. mod . * 16 * ( 2007 ) 1803 . s thirukkanesh and s d maharaj , _ class . quantum grav . _ * 25 * ( 2008 ) 235001 - 1 . r sharma and b s ratanpal , _ int * 22 * ( 2013 ) 1350074 - 1 . m s r delgaty and k lake , _ computer physics communications _ * 115 * ( 1998 ) 395 .	the charged anisotropic star on paraboloidal spacetime is reported by choosing particular form of radial pressure and electric field intensity . the non - singular solution of einstein - maxwell system of equation have been derived and it is shown that model satisfy all the physical plausibility conditions . it is observed that in the absence of electric field intensity , model reduces to particular case of uncharged sharma & ratanpal model . it is also observed that the parameter used in electric field intensity directly effects the mass of the star .
brownian motion of proteins in solution is a subject of recent interest . correlated random walk of proteins within cells may have functional significance in intra - cellular transport and cell migration @xcite . recently , experimentally it has been shown that the brownian motions of proteins within live cells shows bias , which is not observed when the same protein is in the solution@xcite . in this paper we studied the effect of external spin ( magnetic field ) on the random walk of proteins as well as on the sub - cellular streaming pattern . there are reports where effect of magnetic field on diffusion has been studied theoretically . static magnetic field ( smf ) can alter diffusion properties due to occurrence of lorentzian force and maxwell stress . theoretically it has been sown that vortex like motion or constricted phase space ( reduced extent of random motion in space ) is observed upon magnetic field exposure @xcite . the differential diffusional property is the outcome of competition between inertial force and induced lorentzian force . this is the classical view on effect of magnetic field on diffusion . but if the system has an inherent tendency towards correlated random walk , the underlying mechanism is worth exploring from an alternative perspective . + short and long range correlation is an inherent property of ising lattices @xcite . ising model has been used to explain the protein aggregation @xcite , protein folding @xcite , dna - protein interaction @xcite and very recently cancer metastasis @xcite . amino acids in proteins and proteins in solution can both be considered as nodes of a 2d ising lattice at different hierarchical level @xcite . hence , several emerging properties can be explained based on short or long range correlations within and between the lattices . recently , extended 1d ising model has been used to probe the long range correlations within dna . + the present work is inspired by a question what if there are real spin lattice like behavior of proteins in solution . our earlier finding @xcite that tryptophan a building block of proteins show memory of static magnetic field . many of recent interest in quantum mechanical properties of biological molecules also stems from tryptophan distribution @xcite . our experimental design is based on probing for a response to initial external perturbation by a spin perturbing system like static magnetic field ( smf ) . as reports of effect of smf on biomolecules and cells upon long term ( from hours to days ) exposure to smf are numerous @xcite , the evidence of instant response of towards smf or the exact mechanism by which live cells translates the magnetic field and competes thermal energy and retains or memorizes such perturbation are rarely discussed . our experimental approach seems to provide a basis for existence of a real ising lattice like behavior of proteins internally in solution and when they are present in cells . we first show how the zero - field and with field magnetization alters when we subject tryptophan ( trp)a @xmath0-ring containing amino acid to a pre - exposure to static magnetic field while drying the liquid film of the same . while alteration of its fluorescence emission in response to smf was previously reported by our group @xcite , the present work unfolds certain intricate aspects of the magnetic properties of this important ( and universal ) building block of proteins . experiments with squid ( superconducting quantum interference device ) based vsm ( vibrating sample magnetometer ( vsm ) reveals a super paramagnetic behavior of trp . figure 1 shows the zero - field ( solid line ) and with field ( broken line ) magnetization profile of trp at different temperatures.the super paramagnetic nature of the amino acid is confirmed by presence of a neel temperature @xmath1 ( indicated by arrow in figure 1a ) @xcite at which @xmath2 changes sign from positive to negative value . the red and blue profiles compare samples pre - exposed and unexposed to smf . incidentally , the existence of @xmath1 for tryptophan ( @xmath3 ) asserts its super paramagnetic behavior , this being reported for the first time . it may be further noted that blocking temperature for field exposed and unexposed samples are identical i.e. , @xmath4 . now , @xmath5 is a product of magnetic anisotropy and domain volume . as magnetization of the exposed sample ( zero field , red line ) shows a higher value at @xmath1 , a higher magnetic anisotropy is expected in the said case , and this has to be compensated by reduction of magnetic volume as result of the smf exposure . + we performed scattering based imaging using a thin film chamber of three different protein solution before and after magnetic field exposure ( see methods section for experimental details ) . figure 2 shows the differential 3d intensity histogram of the population before and after one and five minutes of magnetic field exposure for bsa ( figure 2a , 2b and 2c ) , ferritin ( figure 2d , 2e and 2f ) and fibrinogen ( figure 2 g , 2h and 2i ) . 3d figures represent intensity histogram of individual particle , which has been imaged . a field dependent change in intensity histogram is clear from figure 2 . but the effect is less pronounced for fibrinogen ( compare figure 2a with 2b , 2d with 2e and 2 g with 2h ) . but after five minutes of field exposure change in distribution pattern can be observed ( see figure 2i ) . + to quantify the change observed in fig . 2,we performed image analysis . we constructed an image plane , which is divided into 10,000 grids ( 100x100 ) , where x - y trajectories of the proteins was plotted . see appendix section for details of the m - code used to construct the image . in fig 3 the panel a , b , c respectively represents the x - y trajectories before exposure after 1 minute exposure and after 5 minute exposure to 0.2 tesla static magnetic field . the panel d represents the rgb representation of the pixels in a b and c. the presence og white pixels indicate that those segments of the xy space where the protein motion is independent of field exposure . the figure rows a , b and c of fig.3 respectively represents studies on bsa , ferritin and fibrinogen . the white pixels are minimal in ferritin and maximal in fibrinogen indicating maximal and minimal magnetic memory in the two respectively . + the fig . 3 further implies that before magnetic field exposure the phase space is isotropic ( space filling ) . a little bit anisotropy in space is observed after 1 minute field exposure and after 5 minutes of exposure the effect become more prominent for bsa and ferritin . for fibrinogen the effect is not so prominent , but a rotation in the direction of the trajectories can be clearly observed . for quantification of the extent anisotropic phase space we calculated an index and assigned as space inhomogeneity index ( sii ) . sii is the ratio of space filled by the proteins and space filled by random distribution of same number of events . table 1 shows the sii values of proteins before and after magnetic field exposure . the values within the fast bracket indicate the respective drift velocity during video acquisition . + figure 4 shows a schematic diagram of the arrangement for confocal microscopic use , in which the instant real time and post smf exposure has been studied . as the proteins in solutions showed the remarkable memory effect similar memory effect could be observed in case of intracellular protein movement . in this case both the instant response and the memory effect was studied . we captured six time lapse images of gfp transfected live hela cells in absence ( image 1 at t=0 minute and image 2 at t=1 minute ) , presence ( image 3 at t=2 minute and image 4 at t=3 minute ) and after withdrawal ( image 5 at t=4 minute and image 6 at t=5 minute ) of the field . the schematics of the magnetic field assisted confocal microscopy and a representative gray scale image of gfp transfected live hela cell is given in appendix section ( see figure s1 and s2 ) . from the six time lapse images we constructed gray scale difference images . control difference image ( see figure 5a ) is constructed by subtraction of image 1 ( t=0 ) from image 2 ( t=1 ) . the magnetic effect is observed by construction of difference image ( see figure 5b ) through subtraction of image 1 ( t=0 ) from image 4 ( t=3 ) . magnetic memory effect is observed by construction of difference image ( see figure 5c ) through subtraction of image 1 ( t=0 ) from image 6 ( t=5 ) . comparison of figure 1a , 1b and 1c clearly shows that magnetic field exposure results in clustered patchy pixel distribution in the image palne . the effect can also be observed even after the withdrawal of the field . it was natural to question whether the sttaic field had any effect on the protein structure . any structural change would lead to changes in the protein walk pattern . our studies using circular dichroism ( cd ) or ans fluorescence showed no structural change in the time scale of the study . this indicated that either such changes are absent or the changes if any had an ultra short duration . the reason for the re - orientation of the walk seemed to be the spin - correlated brownian motion rather than structural change . the long term memory of the walk pattern to such field also implied presence of such correlation rather that structural change in proteins . the spin state of the protein building block thus affected the overall protein motion . + next , the direct evidence of alteration in diffusion property ( random walk pattern ) of proteins by smf comes from image analysis of the acquired video ( see figure 3 ) . the smf induced anisotropic phase space can be compared from table 1 . sii value close to 1 means the space is filled homogenously and lower value of sii indicates inhomogeneity in space . that means , smf induces inhomogeneity in phase space of particle motion ( anisotropic movement ) . from table 1 it is clear that the magnetic field effect is prominent where the drift velocity is much higher . the initial drift velocity ( compare the values within first bracket of table 1 ) in case of bsa and ferritin is very high than fibrinogen . the effect of smf is reduction of drift velocity , which can be observed in naked eye from the video s1 , s2 and s3 . the effect is evidently amplified if the initial drift velocity is high . as we imaged the same sample before and after field exposure there is no change in drift velocity due to change in injection flow . presence of drift velocity ( even under no field condition ) indicates presence of an initial bias in random walk . the particles are moving towards a direction ( biased ) with their intra - particle random motions . application of field imparts a higher lorentzian force that initially perturbs the spin to a greater extent.such exposure breaks the bias and results in an anisotropic streaming of particles in phase space . but , explanation of the memory like effect that is persistence of space anisotropy after field exposure needs to consider smf induced spin coherence in space . the spatial coherence is a inherent property of evolution of ising lattice , where memory is a classical output . the two events that seems to control the spin perturbation induced distribution can be summarized as ( i ) protein shape ( ii ) drift velocity . eventually , the physical ising lattice model explains both the effects . the shape effect is intrinsically linked with dimensionality of the physical ising lattice . a chain like protein structure ( fibrinogen ) may be physically equated to a one dimensional ising lattice . it is known in literature that one dimensional ising lattice is devoid of any magnetic memory and can explain only paramagnetic behavior ( lack of memory ) . in cellular level we also found this type of memory effects . in relaxation image ( image without field exposure subtracted from image after withdrawal of the field , i.e , image 6-image 1 ) we found the retention of the anisotropic subcellular streaming pattern ( see figure 3c ) . that is exposure of magnetic field is memorized by cells in terms of differential streaming pattern . it is well established that the motion of subcellular proteins is biased . not only that , but motor like proteins shows brownian ratchet @xcitelike movement , required for cargo transport , where energy for unidirectional movement ( bias ) comes from atp hydrolysis @xcite . hence , we can say that smf destroys the inherent biased ( correlated ) random walk of sub - cellular proteins . the translation of physical force ( smf ) is altered sub - cellular streaming , which may be the cause of differential effects on cellular level after long term magnetic incubation . as normal and cancer cell can be considered as different ising lattices @xcite , the onset of magnetic memory in live cell can also be considered as an alternative spin state of the same lattice.the differential effect on static magnetic field on normal and cancer cells reported by our group @xcitecan thus be explained . + ising model has been exploited in explaining various binding processes from last four decades one historically important citation being @xcite . the minimal ising lattice , in which the hamiltonian is given by , @xmath6 in the euqtion ( 4 ) @xmath7 can have two values @xmath8 and the first term in the lhs represents the coupling between spins that can exist in absence of a field . the second term involving @xmath9 is the contribution due to bohr magneton that exists only in presence of any external magnetic field b. the superparamagnetic behavior of tryptophan asserts that the ising lattice physically exists for this essential protein building block . the presence of a magnetic domain however does not automatically imply emergence of a magnetic memory as one domensional ising lattice would not exhibit any ferromagnetic memory that would persist in absence of a field . one needs to assume a 2d ising behaviour whose equations in absence of an external magnetic field are given by , @xmath10 where k and l are vertical and horizontal interaction , one representing the side to side ( horizontal ) and the other representing a vertical interaction that may originate from say @xmath11 stacking . such 2d model with periodic boundary condition may be approximated to a globular protein , where for proteins like fibrinogen the behavior would be more like a one dimensional lattice + once we have the insight in the magnetic properties the proteins which can be considered as a decorated lattice containing superparamegnetic domains . we can then look upon the protein walk as a brownian motion of superparamagnetic particles @xcite . while the authors in these work did not consider the memory of an initial magnetic perturbation , their analysis highlights importance of brownian motion of superparamagnetic particles in context of drug targeting and designed magnetic arrays(traps ) . our findings highlight that even proteins can be subjected to magnetically controlled behavior , the memory component of the magnetic contribution depending on the protein shape and dimensionality . it may be noted that simple diamagnetic properties would not suffice to explain the memory aspect . + it also follows from table 1 that the random walk seems to have least spin perturbation memory for fibrinogen , whereas for albumin and ferritin there seems to be a significant random walk shift . the space filling index provides a good estimate of the memory . interestingly , the simple numerical expression can be applied to determine the relative magnetic moments of ferrofluids and other magnetic nanomaterials . + the observation may have important implications in another context . the presence super paramagnetism in the individual amino acid scale and emergence of a ferromagnetic like memory manifested in the random walk in solutions ( and also in protein motion in cells ) implies presence of a temporally stable long term spatial coherence in physiological conditions . in quantum biology some approaches have been sensing weak field @xcite and events triggered by extreme small scale temporal rhythms often termed as quantum beats @xcite.presence of large spatial coherence that is reflected in random walk for a wide variety of proteins , has rarely been discussed . the observation may therefore be contextual with respect to deciphering many cellular communication processes , where very raraely the importance of spatial coherence is discussed . very recently importance of protein dynamics has been emphasized in modification process like ubiquitination and protein - protein interaction . the presence of local spatial coherence may modulate or may open up control of such interactions . the proteins movements _ in vitro _ ( i.e in solution ) and _ in vivo _ ( i.e within the cytoplasm ) show a field memory , this being reported for the first time . mechanistically the behavior is explained by the finding that the building block of proteins , e.g. its aromatic amino acid tryptophan may have super paramagnetic behavior . this inspired a direct exploitation of the ising model , the previous ising model based studies on biomolecular interactions @xcite being primarily analogy based . the authors wish to thank dr . s.chattpadhyay for the squid /vsm experiment at crnn . l - tryptophan , ferritin and fibrinogen were purchased from sigma . the bovine serum albumin ( bsa ) was purchased from srl . millipore water was used to prepare stock solution and dilution . protein solutions were prepared using 100 mm phosphate buffer of ph 7.22 . 20 mg l - tryptophan was dissolved in 10 ml of milliq water . + the permanent neodymium magnets were purchased from rare earth magnetics , india . the field strength was measured using a locally made gaussmeter ( neoequipments , india ) . the prepared 10 ml trytophan solution was divided into two parts and dried using a lyophillizer , one part in absence of magnetic field and other one in presence of 0.2 tesla magnetic field . then the magnetic measurement was performed in vibrational sample magnetometer ( quantum design , mpms 7 ) . both samples were heated to 300k in presence of 100 oesterd field and then cooling up to 4k was performed in presence ( field cooling ) and in absence ( zero field cooling ) of magnetic field . magnetization vs field ( m - h ) study was performed using oscillatory field from -10,000 to + 10,000 oe externally applied magnetic field . nta measurements were performed using nanosight ns300 ( malvern , united kingdom ) system equipped with a high sensitivity cmos camera and 532 nm laser . all measurements were performed at room temperature.the software used for capturing and analyzing the data was the nta 2.3 . 0.1 mg / ml protein solution was imaged before , after 1 minute and after 5 minutes of magnetic field exposure . the video imaging was performed for 1 minute for each condition at 24 frames per second rate . hela cell lines ( passage no : 15 ) was procurred from nccs , india . cells were cultured in dmem , high glucose ( gibco 12800 - 058 ) with 10 percent heat inactivated fbs and the media is supplemented with 1x antibiotic ( penstrep gibco ) . cells were maintained at 5 percent co2 ( carbon di oxide ) environment . for subsequent passages , cells were dislodged using 0.25 percent trypsin edta ( 1x ) ( gibco 25200 ) . hela cells were transfected by pegfp - n3 plasmids using calcium phosphate transfection protocol as described previously @xcite . time lapse confocal imaging was performed in olympus ( fv1000 ) microscope . six time lapse images were acquired at 1 minute time interval . first and second image were captured in absence of magnetic field . third and fourth image were captured in presence of magnetic field . the fifth and sixth image were captured after withdrawal of the magnetic field . the schematic illustration of the magnetic field assisted microscopy is provided in the supplementary section . + m - code 42 ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1 '' '' @noop [ 0]secondoftwo sanitize@url [ 0 ] + 12$12 & 12#1212_12%12 @startlink[1 ] @endlink[0 ] @bib@innerbibempty @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) in @noop _ _ , vol . ( , ) pp . @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) in @noop _ _ , vol . ( , ) pp . @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop ( ) @noop * * , ( ) @noop * * , ( )	superparamagnetism of tryptophan implying the presence of magnetic domain is reported . the observation helps us to conceive assembly of proteins as a physical lattice gas with multidimensional ising character , each lattice points assuming discrete spin states . when magnetic field is applied the equilibrium is lost and the population density of one spin state increases ( unidirectional alignment ) , resulting in net magnetization . spatial coherence between the identical spin states further imparts a ferromagnetic memory . this effect is observed using direct nanoscale video imaging . out of the three proteins ferritin serum albumin and fibrinogen , fibrinogen showed an attenuated response , the protein being essentially one dimensional . eventually , ising lattice is capable of showing ferromagnetic memory only when it has a higher dimensional character . the study highlights possible presence of long range spatial coherence at physiological condition and a plausible microscopic origin of the same . . pacs numbers : : 61.48.de,82.60.cx,82.70.uv,82.33.nq keywords : : * * carbon nanomaterials , enthalpy , coupling , surface topology .
unlike four dimensional einstein gravity , which allows existence of black holes , higher dimensional theories , such as the brane - world scenarios and string theory , allow existence of a number of `` black '' objects : higher dimensional black holes , black strings and branes , black rings and saturns and others . in higher than four dimensions we lack the uniqueness theorem , so that stability may be the criteria which will select physical solutions among this variety of solutions . up to now , we know that higher dimensional reissner - nordstrm - de sitter black holes are stable @xcite in the einstein gravity . on the contrary , black holes in gauss - bonnet ( gb ) gravity are unstable for large gb coupling for @xmath0 @xcite , where @xmath1 is the total number of space - time dimensions . kaluza - klein black holes with squashed horizon are stable against lowest zero mode perturbations @xcite . unlike , kaluza - klein black holes , the black string metric is a solution of the einstein equations in five or higher dimensional gravity that has a factorized form consisting of the tangherlini black hole and an extra flat dimension @xcite . according to the brane - world scenarios , if the matter localised on the brane undergoes gravitational collapse , a black hole with the horizon extended to the transverse extra direction will form . this object looks like a black hole on the brane , but is , in fact , a black string in the full d - dimensional theory . it is well known that such black strings suffer from the so - called gregory - laflamme instability , which is the long wavelength gravitational instability of the scalar type of the metric perturbations @xcite , @xcite . the gregory - laflamme instability has been intensively studied during the recent decade @xcite and the threshold values of the wave vector @xmath2 at which the instability appears are known @xcite . in the present paper we are aimed at studying the evolution of linear perturbations of d - dimensional black strings in time and frequency domains . this task is motivated mainly by the two reasons : first to realize what happens on the edge of instability of black strings and how the perturbations will develop in time . second , the so - called quasinormal modes of a stable black string might be an observational characteristic for the future large hadron collider experiments , if such objects as black strings exist . the latter needs some more explanation . in this research , we shall show that , if a stable d - dimensional black string is gravitationally perturbed , it will undergo damped oscillations , called quasinormal ringing , similar to that of a black hole @xcite . at asymptotically late time , this quasinormal ringing goes over into the power - law tails . the quasinormal modes and asymptotic tails are very well studied for d - dimensional black holes @xcite and for black holes localised on the brane @xcite . we find here that the quasinormal ringing of black strings has a number of differences from that of d - dimensional black holes @xcite , especially when approaching the edge of instability . in particular , we find a numerical evidence that in the threshold point of instability , the static solution of the wave equation is dominant . in the region of the stability , for @xmath2 slightly larger than the threshold value , we see modes with tiny oscillation frequencies and damping rates . in the region of the gregory - laflamme instability , for @xmath2 slightly smaller than threshold value , we observe tiny oscillations and instability growth . we also find the level crossing of imaginary part of quasinormal modes between the fundamental mode and the first overtone mode . the paper is organized as follows . sec ii gives basic formulas for the black string perturbations and the wave equation for the scalar type of gravitational perturbations . sec iii analyses quasinormal modes of black strings in frequency domain with the help of the frobenius method , and in time domain by the gundlah - price - pullin method . we also discuss features of the gregory - laflamme instability using the obtained results . the final section is devoted to the conclusion . in this section , we shall briefly review the results of the paper @xcite , where the wave equation for the scalar type of gravitational perturbations were obtained . this wave equation will be our starting point for numerical investigation . for the static string in @xmath3 space - time dimensions , the background metric can be written as @xmath4 where @xmath5 and @xmath6 is the metric on a unit @xmath7-sphere . various properties of black string have been extensively studied in recent years and we refer a reader to the papers @xcite-@xcite for more detailed information on black strings . the @xmath8-direction is periodically identified by the relation @xmath9 . we study perturbations of an @xmath7-spherically symmetric solution with the killing vector in @xmath8-direction . therefore , we can write perturbations in the form @xmath10 the perturbed vacuum einstein equations have the form @xmath11 the perturbations can be reduced to the form , where the only non - vanishing components of @xmath12 are @xmath13 @xmath14 the linearized einstein equations give a set of coupled equations determining the four radial profiles above . however , we may eliminate @xmath15 , @xmath16 and @xmath17 from these equations in order to produce a single second order equation for @xmath18 : @xmath19 where @xmath20 defining @xmath21 we can reduce the equation ( [ hovdebomyers ] ) to the wave - like equation @xmath22 where @xmath23 is the tortoise coordinate . here , the effective potential @xmath24 is given by @xmath25 where @xmath26 the above effective potential does not vanish at asymptotic infinity but has an effective `` mass '' term , containing @xmath2 , at the spatial infinity . first of all , let us briefly describe the two methods which we used here : the frobenius method ( frequency domain ) and the gundlach - price - pullin method ( time domain ) . in time domain , we study the black string ringing using a numerical characteristic integration method @xcite , that uses the light - cone variables @xmath27 and @xmath28 . in the characteristic initial value problem , initial data are specified on the two null surfaces @xmath29 and @xmath30 . the discretization scheme we used , is @xmath31 where we have used the following definitions for the points : @xmath32 , @xmath33 , @xmath34 and @xmath35 . in frequency domain we used the well - known frobenius method @xcite . in order to study the qn spectrum in frequency domain , we have separated time and radial coordinates in ( [ hovdebomyers ] ) @xmath36 here @xmath37 satisfies the quasinormal mode boundary conditions , which are purely ingoing wave at the event horizon and purely outgoing wave at the spatial infinity . thus , the appropriate frobenius series are @xmath38 where @xmath39 for @xmath40 and @xmath41 for @xmath42 . it is crucial that @xmath43 must be regular at the event horizon and at the spatial infinity and can be expanded as @xmath44 substituting ( [ frobenius ] ) into ( [ hovdebomyers ] ) , we have found that the coefficients @xmath45 satisfy a @xmath46-term recurrence relation . we found the coefficients of the recurrence relation , and then we obtained the equation with the infinite continued fraction , which is algebraic equation with respect to the qn frequency @xmath47 . numerical solutions of this algebraic equation give us the qn spectrum . .fundamental mode ( @xmath48 ) found by the time - domain integration , first ( @xmath49 ) and second ( @xmath50 ) overtones of spherically symmetric black string perturbations ( @xmath40 ) found by the frobenius method . as @xmath2 grows , the first overtone decreases its imaginary part , becoming the fundamental mode , reaching purely real frequency ( quasi - resonance ) at @xmath51 and then disappearing . the second overtone , as well as the fundamental purely imaginary mode , increases its damping rate with @xmath2 . [ cols="^,^,^",options="header " , ] in table iii , we listed dominant qnms of spherically symmetric black string perturbations in the case of @xmath52 . we can see the level crossing of imaginary part of quasinormal modes between the fundamental mode and the first overtone mode . the real part of the second mode ( @xmath49 in tables i , iii ) asymptotes to @xmath2 at large @xmath2 , while the imaginary part monotonically decreases when @xmath2 is increasing . the third mode does not asymptotes to @xmath2 , but has monotonically decreasing real and imaginary parts as can be seen from fig . 1 . because of the risk to `` fall '' into another overtone , in order to obtain the higher overtones , we had to start from the d - dimensional black holes with @xmath53 , for which the qn frequencies are known @xcite , and then to `` move '' towards higher @xmath2 in the frobenius method with a very small step ( see fig . 1 ) . @xmath54 ( red , top ) , @xmath52 ( orange ) , @xmath55 ( green ) , @xmath56 ( blue , bottom ) . late - time decay of perturbations for @xmath57 is @xmath58 . for lower @xmath59 the law of decay is different : @xmath60 for @xmath40 and @xmath61 for @xmath54 . ] an essential advantage of the time domain method in comparison with the frobenius method is that we do not have any decreasing of the convergence or loss of accuracy when approaching the point of instability . therefore our time domain method is more complete than the frequency one , at least for the dominant mode , which can always be extracted from the time domain picture . in fig . 2 , one can see examples of time domain profiles for various @xmath59 and a fixed @xmath2 . there one can see that for @xmath62 the intermediate late time asymptotic is power law like @xmath63 while for other @xmath59 , the asymptotics are @xmath64 @xmath65 let us note that this asymptotic apparently should be considered as intermediate . they are expected to go over into other power law ones at very late times , as it takes place for massive fields in general @xcite , @xcite , @xcite , @xcite . let us note , that @xmath2 plays the role of the effective mass . at asymptotically late time we observe power - law damped tails , which have oscillation frequency equal to @xmath2 , resembling asymptotical behavior of massive fields near schwarzschild black holes . the first overtone s behavior is qualitatively similar to that of the fundamental mode for massive fields of higher - dimensional schwarzschild black holes @xcite : we can see long - lived oscillations , which can be infinitely long lived modes called the quasi - resonances @xcite . the analytical explanation of existence of the quasi - resonances was found in @xcite . * for @xmath66 ( @xmath40 ) , as @xmath2 grows , the imaginary part of the first overtone quickly decreases and vanishes for some threshold value of @xmath2 , while its real part stays smaller than the threshold value ( see table [ n=1.table ] ) . after the threshold value of @xmath2 is reached , the first overtone `` disappears '' . * for @xmath67 ( @xmath68 ) , the imaginary part of the first overtone becomes small for large @xmath2 , while the real part asymptotically approaches @xmath2 ( see fig . [ n=2.fig ] ) . even though the first overtone of the spherically symmetric black strings behaves similarly to the fundamental mode of massive fields near higher - dimensional schwarzschild black holes , the other modes have completely different behavior . the fundamental mode of a black string perturbation is purely imaginary . it grows for small values of @xmath2 , leading to instability of the black string . this behavior is common for the unstable modes . indeed , let us multiply the equation ( [ wavelike ] ) by the complex conjugated function @xmath69 and assume that the dependance on time is @xmath70 . let us study the integral of the obtained equation @xmath71 integration of the first term by parts gives @xmath72 taking into account the boundary conditions ( [ frobenius ] ) , we find that the imaginary part of the integral is @xmath73 since the sign of @xmath74 coincides with the sign of @xmath75 , the non - zero real part of the quasi - normal frequency implies that the imaginary part is negative . therefore , the unstable modes ( @xmath76 ) must have zero real part . in other words , _ unstable modes can not be oscillating_. @xmath77 ( magenta , top ) , @xmath78 ( red ) , @xmath79 ( orange ) , @xmath80 ( green ) , @xmath81 ( blue , bottom ) . we can see two concurrent modes : for large @xmath2 the oscillating one dominates , near the critical value of @xmath2 the dominant mode does not oscillate ( looks like exponential tail ) , for unstable values of @xmath2 the dominant mode grows . the plot is logarithmic , so that straight lines correspond to an exponential decay . ] + in the stability region , the fundamental mode is also purely imaginary , but damped mode , whose damping rate grows quickly with @xmath2 . because of its quick growth , this mode can not be considered as the fundamental one for large @xmath2 . in fact , for larger @xmath2 ( see tables i - iii ) , the first overtone turns out to be the fundamental mode ( the mode with the largest lifetime ) . the real part of the second overtone decreases , as @xmath2 grows , and reaches zero for some value of @xmath2 , while the imaginary part remains negative . let us now look at the fig.3 . at moderately large values of @xmath2 , sufficiently far from instability , the profile has the same form as that for massive fields , yet , when approaching the instability point the real oscillation frequency ( @xmath82 ) and the decay rate ( @xmath83 ) decrease considerably . after crossing the instability point we observe , that starting from some tiny values , @xmath84 are slowly increasing ( while @xmath82 is still zero for the fundamental mode and tiny oscillations , observed in the time domain , come from the next decayed mode ) . therefore we conclude , that the there is some _ static solution _ @xmath85 of the wave equation ( 4 ) , which shows itself _ exactly in the threshold point of instability_. we would say that this picture of instability is natural , if the instability develops on the fundamental mode . however , if instability occurs at higher multipoles @xmath86 , as it takes place for instance in the gauss - bonnet theory @xcite , the picture of instability is quite different : growing modes appear only after rather long period of decaying oscillations ( see @xcite ) . note also , that here we confirmed the threshold values of @xmath2 found in @xcite with a very good accuracy by the time domain integration ( see for instance fig . 3 for @xmath40 ) . thus the threshold values are : @xmath87 for @xmath40 , @xmath88 for @xmath54 , @xmath89 for @xmath52 , @xmath90 for @xmath55 . finally , in fig . 4 , we can see the region of the profiles where the period of the quasinormal ringing goes over into the power - law tail behavior . close to the critical point , there exists a period where the oscillation ceases . this is because the pure damping mode becomes the fundamental mode near the critical point . we have numerically studied the gregory - laflamme instability through quasinormal modes . let us stress three main results obtained here : \1 ) we have found the quasinormal modes and late - time tails for scalar type of gravitational perturbations of d - dimensional black strings for various @xmath1 , that is for the type of perturbations where the gregory - laflamme instability forms in the long wavelength regime . \3 ) near the instability point ( in @xmath2 ) the fundamental mode is pure imaginary ( non - oscillating ) , and , as @xmath2 is increasing , the lifetime of the second mode is increasing , so that at some moderate @xmath2 both modes are dominating at the late time of the ringing . at larger @xmath2 , the dominance goes over to the second ( oscillating ) mode , as to the longer lived one . our research could be improved in a number of ways . first of all , one could compute qnms for higher multipole numbers , starting from the effective potential derived in @xcite and also for other types of gravitational perturbation . though vector and tensor types of perturbations do not contain instabilities , such investigation would give us complete data on qnms and evolution of gravitational perturbations . the main limitation of our analysis is that we can not say what happen with unstable black string some time since the moment of initial perturbations : the perturbations will grow and become large , so that the linear approximation will not be valid anymore . however , it is beyond the scope of this paper . r. a. k. was supported by the japan society for the promotion of science ( jsps ) , japan . k.m . was supported by jsps grant - in - aid for scientific research no . 19 @xmath92 3715 . j.s . was supported by the japan - u.k . research cooperative program , grant - in - aid for scientific research fund of the ministry of education , science and culture of japan no.18540262 and no.17340075 . a. z. was supported by _ fundao de amparo pesquisa do estado de so paulo ( fapesp ) _ , brazil . 99 h. kodama and a. ishibashi , prog . theor . phys . * 111 * , 29 ( 2004 ) [ arxiv : hep - th/0308128 ] ; a. ishibashi and h. kodama , prog . phys . * 110 * , 901 ( 2003 ) [ arxiv : hep - th/0305185 ] ; r. a. konoplya and a. zhidenko , nucl . b * 777 * , 182 ( 2007 ) [ arxiv : hep - th/0703231 ] . r. a. konoplya and a. zhidenko , phys . d * 77 * , 104004 ( 2008 ) arxiv:0802.0267 [ hep - th ] ; m. beroiz , g. dotti and r. j. gleiser , phys . d * 76 * , 024012 ( 2007 ) [ arxiv : hep - th/0703074 ] . h. ishihara , m. kimura , r. a. konoplya , k. murata , j. soda and a. zhidenko , phys . rev . d * 77 * , 084019 ( 2008 ) [ arxiv:0802.0655 [ hep - th ] ] ; m. kimura , k. murata , h. ishihara and j. soda , phys . d * 77 * , 064015 ( 2008 ) [ arxiv:0712.4202 [ hep - th ] ] . r. gregory and r. laflamme , phys . . lett . * 70 * , 2837 ( 1993 ) [ arxiv : hep - th/9301052 ] . r. gregory and r. laflamme , nucl . b * 428 * , 399 ( 1994 ) [ arxiv : hep - th/9404071 ] . t. harmark , v. niarchos and n. a. obers , class . * 24 * , r1 ( 2007 ) [ arxiv : hep - th/0701022 ] . j. l. hovdebo and r. c. myers , phys . d * 73 * , 084013 ( 2006 ) [ arxiv : hep - th/0601079 ] . k. kokkotas , b. schmidt , liv . 02 , 2 , 1999 r. a. konoplya , phys . d * 68 * , 124017 ( 2003 ) [ arxiv : hep - th/0309030 ] . r. konoplya , phys . d * 71 * , 024038 ( 2005 ) [ arxiv : hep - th/0410057 ] ; v. cardoso , j. p. s. lemos and s. yoshida , jhep * 0312 * ( 2003 ) 041 [ arxiv : hep - th/0311260 ] ; h. t. cho , a. s. cornell , j. doukas and w. naylor , phys . d * 75 * ( 2007 ) 104005 [ arxiv : hep - th/0701193 ] ; r. g. daghigh , g. kunstatter and j. ziprick , class . * 24 * ( 2007 ) 1981 [ arxiv : gr - qc/0611139 ] ; f. w. shu and y. g. shen , jhep * 0608 * ( 2006 ) 087 [ arxiv : hep - th/0605128 ] ; r. a. konoplya , phys . d * 68 * , 024018 ( 2003 ) [ arxiv : gr - qc/0303052 ] ; r. a. konoplya , phys . d * 66 * , 044009 ( 2002 ) [ arxiv : hep - th/0205142 ] ; r. g. daghigh , g. kunstatter , d. ostapchuk and v. bagnulo , class . * 23 * ( 2006 ) 5101 [ arxiv : gr - qc/0604073 ] ; r. a. konoplya and e. abdalla , phys . rev . d * 71 * , 084015 ( 2005 ) [ arxiv : hep - th/0503029 ] ; e. abdalla , r. a. konoplya and c. molina , phys . d * 72 * , 084006 ( 2005 ) [ arxiv : hep - th/0507100 ] ; a. lopez - ortega , arxiv:0706.2933 [ gr - qc ] ; c. molina , phys . d * 68 * ( 2003 ) 064007 [ arxiv : gr - qc/0304053 ] . p. kanti and r. a. konoplya , phys . d * 73 * , 044002 ( 2006 ) [ arxiv : hep - th/0512257 ] ; p. kanti , r. a. konoplya and a. zhidenko , phys . d * 74 * , 064008 ( 2006 ) [ arxiv : gr - qc/0607048 ] ; m. nozawa and t. kobayashi , arxiv:0803.3317 [ hep - th ] ; u. a. al - binni and g. siopsis , phys . d * 76 * ( 2007 ) 104031 [ arxiv:0708.3363 [ hep - th ] ] ; l. h. liu and b. wang , arxiv:0803.0455 [ hep - th ] . y. kurita and h. ishihara , class . grav . * 25 * , 085006 ( 2008 ) [ arxiv:0801.2842 [ hep - th ] ] . y. brihaye , t. delsate and e. radu , phys . lett . b * 662 * , 264 ( 2008 ) [ arxiv:0710.4034 [ hep - th ] ] . @xcite b. kleihaus and j. kunz , arxiv:0710.1726 [ hep - th ] . @xcite a. bernamonti , m. m. caldarelli , d. klemm , r. olea , c. sieg and e. zorzan , jhep * 0801 * , 061 ( 2008 ) [ arxiv:0708.2402 [ hep - th ] ] . y. brihaye and t. delsate , class . quant . * 24 * , 4691 ( 2007 ) [ arxiv : gr - qc/0703146 ] . k. murata , j. soda and s. kanno , phys . rev . d * 75 * , 104017 ( 2007 ) [ arxiv : gr - qc/0703029 ] ; k. murata , j. soda and s. kanno , arxiv : gr - qc/0701137 . y. brihaye , b. hartmann and e. radu , phys . d * 72 * , 104008 ( 2005 ) [ arxiv : hep - th/0508028 ] . a. tomimatsu , phys . d * 71 * , 124044 ( 2005 ) [ arxiv : gr - qc/0506120 ] . t. kobayashi and t. tanaka , phys . d * 71 * , 084005 ( 2005 ) [ arxiv : gr - qc/0412139 ] . v. cardoso and j. p. s. lemos , phys . b * 621 * , 219 ( 2005 ) [ arxiv : hep - th/0412078 ] . @xcite b. kol , phys . rept . * 422 * , 119 ( 2006 ) [ arxiv : hep - th/0411240 ] . m. i. park , class . * 22 * , 2607 ( 2005 ) [ arxiv : hep - th/0405045 ] . t. tamaki , s. kanno and j. soda , phys . d * 69 * , 024010 ( 2004 ) [ arxiv : hep - th/0307278 ] . s. kanno and j. soda , class . * 21 * , 1915 ( 2004 ) [ arxiv : gr - qc/0311074 ] . s. kanno and j. soda , arxiv : gr - qc/0312107 . c. gundlach , r. h. price and j. pullin , phys . d * 49 * , 883 ( 1994 ) [ arxiv : gr - qc/9307009 ] . e. w. leaver , proc . a * 402 * , 285 ( 1985 ) . l. e. simone and c. m. will , class . * 9 * ( 1992 ) 963 . r. a. konoplya , phys . b * 550 * , 117 ( 2002 ) [ arxiv : gr - qc/0210105 ] a. zhidenko , phys . d * 74 * , 064017 ( 2006 ) [ arxiv : gr - qc/0607133 ] . a. ohashi and m. a. sakagami , class . * 21 * ( 2004 ) 3973 [ arxiv : gr - qc/0407009 ] r. a. konoplya and a. v. zhidenko , phys . b * 609 * , 377 ( 2005 ) [ arxiv : gr - qc/0411059 ] ; r. a. konoplya and a. zhidenko , phys . d * 73 * , 124040 ( 2006 ) [ arxiv : gr - qc/0605013 ] . h. kudoh , phys . d * 73 * , 104034 ( 2006 ) [ arxiv : hep - th/0602001 ] .	we study evolution of gravitational perturbations of black strings . it is well known that for all wavenumber less than some threshold value , the black string is unstable against scalar type of gravitational perturbations , which is named the gregory - laflamme instability . using numerical methods , we find the quasinormal modes and time - domain profiles of the black string perturbations in the stable sector and also show the appearance of the gregory - laflamme instability in the time domain . the dependence of the black string quasinormal spectrum and late time tails on such parameters as the wave vector and the number of extra dimensions is discussed . there is a numerical evidence that in the threshold point of instability the static solution of the wave equation is dominant . for wavenumbers slightly larger than the threshold value , in the region of stability , we see tiny oscillations with very small damping rate . while , for wavenumbers slightly smaller than the threshold value , in the region of the gregory - laflamme instability , we observe tiny oscillations with very small growth rate . we also find the level crossing of imaginary part of quasinormal modes between the fundamental mode and the first overtone mode , which accounts for the peculiar time domain profiles .
in @xcite the authors make some comments about their previous work @xcite and also reply to certain issues that have been raised by some authors @xcite concerning the nature of the matter distribution and the asymptotic behaviour of the spacetime in their model . in particular , @xcite pointed out that the use of a @xmath1 function in the solution of the field equations introduces an additional disk on the @xmath0 plane and later @xcite showed that the disk was made of exotic matter . consequently , @xcite examine the question in several ways and present some arguments to dismiss the existence of a physical mass layer present on the @xmath0 plane . we would like to notice some inconsistencies in their paper . the first is related to the expressions used for the surface density of the singular disk , and is presented in sec . [ sec_disk ] . in sec . [ sec_smooth ] we question the smoothing procedure discussed in section 4 of @xcite . we find that one can not eliminate disk singularities , moreover one introduces new singularities that make the model unrealistic . in @xcite we calculate the distributional energy - momentum tensor @xmath2 due to the introduction of an absolute value of @xmath3 in the metric functions . the resulting non - zero components read @xcite @xmath4 where all quantities are evaluated on @xmath0 . since @xmath2 is non - diagonal , in order to have obtain the physical variables of the disk , we need to put the energy - momentum tensor in its canonical form . to do that we solve the eigenvalue problem : @xmath5 . we thus find @xmath6 and using eq . ( [ eq_qtt])([eq_qphiphi ] ) result in @xmath7 as the discriminant is always negative , the eigenvalues and corresponding eigenfunctions are complex conjugate . if @xmath8 and @xmath9 denote the timelike and spacelike real eigenvectors , the canonical form of the energy - momentum tensor is @xmath10 where @xmath11 is the surface density , @xmath12 denote the azimuthal stresses and @xmath13 is the heat flow in the azimuthal direction @xcite . thus the expression for the surface density , to order @xmath14 , is given by @xmath15 , or , using the relation @xmath16 we get @xmath17 on the other hand , @xcite take following expressions for the surface energy density ( equations ( 15 ) and ( 16 ) of the paper ) @xmath18 which they integrate over the surface and compare with the volume integral of their continuous mass density distributions . but eq . ( [ eq_sigma2 ] ) is only the @xmath19 component eq . ( [ eq_qtt ] ) to order @xmath14 . due to the non - diagonal form of the energy - momentum tensor , the @xmath19 component solely does not determine the surface density , but there is also a contribution from the @xmath20 component eq . ( [ eq_qphiphi ] ) _ even to order _ @xmath14 . thus , the correct expression for @xmath21 that should be used is eq . ( [ eq_sigma ] ) . the integration of this equation over the surface would result in half of the value for the mass derived from the volume integral of the continuous mass distribution , since the authors themselves comment on footnote 7 that the two terms in eq . ( [ eq_sigma2 ] ) contribute equally . it is important to stress that the physical variables of the mass layer on the @xmath0 plane are not directly given by the principal diagonal terms of the distributional energy - momentum tensor , since the non - diagonal terms are non - negligible . this is further an example of how the non - linearity of general relativity can manifest even at newtonian level . another approach used in @xcite to examine the presence of a singular mass layer on the symmetry plane was to smooth the solution over an interval that includes the @xmath0 plane . this was achieved by the choice of @xmath22 functions to span the symmetry plane in the interval @xmath23 . for @xmath24 , the original solution with exponentials was used . this requires that the functions @xmath25 _ and _ also @xmath26 match at @xmath27 . if this last condition is not satisfied new matter is added . we note that this is also true for the newtonian gravitational potential , discontinuity of the first derivatives `` add matter '' whose density can be computed via poisson s equation . we argue that the above mentioned matching can not be done without adding new matter . and impose the conditions @xmath29 and @xmath30 . using eq . ( [ eq_n1])([eq_n2 ] ) we obtain @xmath31=0 , \label{eq_match1}\\ & \sum_n k_n^2rj_1(k_nr ) \left [ -c_{2n}\sinh(k_nz_0)-c_{1n}e^{-k_nz_0 } \right]=0 . \label{eq_match2}\end{aligned}\ ] ] since the bessel functions are linearly independent , the terms in brackets must vanish identically . subtracting eq . ( [ eq_match1 ] ) from eq . ( [ eq_match2 ] ) results in @xmath32 , which is only satisfied in the real domain if all @xmath33 . thus it is not possible to match @xmath25 and @xmath26 simultaneously . if we demand continuity of @xmath34 , @xmath35 and @xmath36 at @xmath27 we obtain the following conditions : @xmath37 on the other hand , demanding continuity of @xmath38 , @xmath39 and @xmath40 at @xmath27 results in @xmath41 fig . [ fig1](a ) sketches what happens if we impose continuity of the functions and fig . [ fig1](b ) if we impose continuity of the derivatives . the only way the functions eq . ( [ eq_n1])([eq_n3 ] ) _ and _ their derivatives could be matched is if an extra set of constants were inserted into eq . ( [ eq_n2 ] ) @xmath42 but then this would not be a solution of @xmath43 assuming conditions eq . ( [ eq_cond ] ) hold , the jump of the derivatives of eq . ( [ eq_n1])([eq_n3 ] ) with respect to @xmath3 evaluated at @xmath27 are given by @xmath44 the same kind of descontinuities in the derivatives also appear when a @xmath1 is introduced in the solution . thus they introduce additional layers of matter now located on the @xmath45 planes that makes the solution unrealistic . 9 f. i. cooperstock and s. tieu , preprint : astro - ph/0512048 . f. i. cooperstock and s. tieu , preprint : astro - ph/0507619 . m. korzyski , preprint : astro - ph/0508377 . d. vogt and p. s. letelier , preprint : astro - ph/0510750 . d. garfinkle , preprint : gr - qc/0511082 . c. mller , _ the theory of relativity _ , oxford university press , 1972 .	in this comment we question some arguments presented in astro - ph/0512048 to refuse the presence of an singular mass surface layer . in particular , incorrect expressions are used for the disk s surface mass density . we also point out that the procedure of removing the descontinuity on the @xmath0 plane with a region of continuous density gradient generates other two regions of descontinuities with singular mass surface layers making the model unrealistic .
it is a pleasure to contribute to this volume in honor of yoseph imry s 70th birthday . we know joe as a scientist , teacher , co - worker and friend . djs met joe in the early seventies when joe first came to ucsb . he remembers how they often talked about how one might gain insight into the basic physics of a macroscopic system from calculations on small subsystems and how these subsystems had their own interesting features . all of us are eager to recognize how much we have learned from joe s work , over the years concerning the intrinsic , subtle and beautiful physics one finds in the mesoscopic world . this work was supported in part by the national science foundation under the grant phy05 - 51164 at the kitp . djs acknowledges the center for nanophase materials science at ornl , which is sponsored by the division of scientific user facilities , u.s . eb was supported by the nsf under grants dmr-0705472 and dmr-0757145 at harvard . sak was supported , in part , by the nsf grant number dmr-0758356 at stanford . emery , s.a . kivelson and h.q . lin , _ phys . lett . _ * 64 * , 475 ( 1990 ) . anderson , _ science _ * 235 * , 1196 ( 1987 ) . c. gros , r. joynt and t.m . z. phys . b _ * 68 * , 425 ( 1987 ) . ruckenstein , p.j . hirschfeld and j. appel , _ phys . b _ * 36 * , 857 ( 1987 ) . g. kotliar and j. liu , _ phys . rev . _ * 38 * , 5142 ( 1988 ) . bickers , d.j . scalapino and r.t . scalettar , _ int . j. mod . phys . _ * b1 * , 687 ( 1987 ) . scalapino and s.a . trugman , _ philos . mag . b _ * 74 * , 607 ( 1996 ) . carlson , v.j . emery , s.a . kivelson and d. orgad , chap . 21 , superconductivity * vol . 1 * , eds . bennemann and j.b . ketterson ( springer , 2006 ) . e. dagatto , j. riera and d. scalapino , _ phys . b _ * 45 * , 5744 ( 1992 ) . for a review see e. dagatto and t.m . rice , _ science _ * 271 * , 618 ( 1996 ) . s. chakravarty , _ phys . lett . _ * 77 * , 4446 ( 1996 ) . a. luther and v.j . emery , _ phys . lett . _ * 33 * , 589 ( 1974 ) . feiguin , s.r . white , d.j . scalapino and i. affleck , _ phys . lett . _ * 101 * , 217001 ( 2008 ) scalapino , y. imry and p. pincus , _ phys . b _ * 11 * , 2042 ( 1975 ) . h. tsunetsugu , m. troyer and t.m . rice , _ phys . rev . b _ * 49 * , 16078 ( 1994 ) . e. arrigoni , e. fradkin and s.a . kivelson , _ phys . b _ * 69 * , 214519 ( 2004 ) . e. altman and a. auerbach , _ phys . b _ * 65,104508 ( 2002 ) . h. yao , w - f . tsai and s. a. kivelson , _ phys . b _ 76 * , 161104 ( 2007 ) . e. berg , s.a . kivelson and d.j . scalapino , _ new j. phys . _ * 11 * , 085007 ( 2009 ) . k. kuroki , s. onari , r. arita , h. usui , y. tanaka , h. kontani and h. aoki , _ phys . rev . lett . _ * 101 * 087004 ( 2008 ) . s. graser , t.a . maier , p.j . hirschfeld and d.j . scalapino , _ new j. phys . _ * 11 * 025016 ( 2009 ) . f. wang , h. zhai , y. ran , a. vishwanath , and d .- h . lee , _ phys . * 102 * , 47005 ( 2009 ) . white , _ phys . lett . _ * 69 * , 2863 ( 1992 ) . c. cao , p.j . hirschfeld and h .- cheng , _ phys . b _ * 77 * , 220506(r ) ( 2008 ) . i.i . mazin , d.j . singh , m.d . johannes and m.h . du , _ phys . lett . _ * 101 * , 057003 ( 2008 ) . noack , n. bulut , d.j . scalapino and m.g . zacher , _ phys . b _ * 56 * , 7162 - 7166 ( 1997 ) . y. imry , _ introduction to mesoscopic physics _ , oxford university press ( 1997 ) . rice , s. haas , m. sigrist and f. zhang , _ phys . b _ * 56 * , 6569 ( 1997 ) . white and d.j . scalapino , _ phys . b _ * 57 * , 3031 ( 1998 ) . e. berg , e. fradkin , s.a . kivelson and j.m . tranquada , cond - mat / arxiv:0901.4826 .	mesoscopic physics concerns itself with systems which are intermediate between a single atom and a bulk solid . besides the many intrinsically interesting properties of mesoscopic systems , they can also provide physical insight into the physics of bulk systems . here we discuss three examples of this from the field of high temperature superconductivity . many interesting electronic materials , especially those that exhibit high temperature superconductivity , lie in an intermediate coupling regime where the strength of the interaction is comparable to the electron bandwidth . the absence of any small parameter makes analytic approaches to such problems difficult . conversely , in the absence of a small parameter there is every reason to expect the correlation lengths associated with any form of electronic order to be of order 1 in units of the lattice constant ( barring an accident which places the system close to a critical point ) . thus , with some sensible analysis ( and with a little bit of luck ) , the properties of the system in the thermodynamic limit may be apparent in the properties of mesoscopic systems , even systems that are small enough that they can be studied by essentially exact numerical methods . here we see how three such studies have provided insight into the high @xmath0 problem . following the discovery of the high @xmath0 cuprates there were various suggestions of ways in which the strongly repulsive hubbard interaction , @xmath1 , between two electrons on the same site could ( paradoxically ) produce high temperature pairing in a doped antiferromagnetic insulator . numerous studies began with the @xmath2 model on a square lattice , which can be thought of as the strong coupling limit of the hubbard model . here it is assumed that @xmath1 is sufficiently large to prevent double occupancy of any site , leaving a one - electron near - neighbor hopping term @xmath3 and an exchange coupling @xmath4 . an early argument for pairing was based upon the observation that if one adds two holes to the half - filled ( one - electron per site ) system , then eight exchange @xmath4 bonds are broken if the holes were well separated . however , if the two holes are placed on near neighbor sites , only seven exchange @xmath4 bonds are broken . thus there is an effective near - neighbor attraction between the holes . this picture however was soon seen as more applicable to the phase separation regime @xcite . alternatively , in the context of the resonance - valence - bond @xcite approach , variational calculations using gutzwiller projected wavefunctions @xcite and auxiliary - boson meanfield @xcite calculations found a superconducting state with @xmath5-wave symmetry in the @xmath2 model . from a more weak - coupling perspective , the idea of spin - fluctuation exchange mediated pairing near an antiferromagnetic instability @xcite was also found to lead to @xmath5-wave pairing due to the increasingly positive strength of the pairing interaction at large momentum transfer . however , none of these approaches gave a simple , crisp real space picture , especially one that makes clear why @xmath5-wave rather than extended @xmath6-wave symmetry is preferred . to address this , trugman and one of the authors @xcite decided to imagine that a 4-site plaquette was extracted from the lattice . the undoped " groundstate of the @xmath2 model on a plaquette with 4 electrons is a singlet having a wave function @xmath7 here , @xmath8 creates a singlet pair on sites @xmath9 and we have numbered the sites of the plaquette in a clockwise manner . this state is odd under a @xmath10 rotation . the two - electron groundstate @xmath11 has spin 0 and is invariant under a @xmath10 rotation . therefore the pairfield annihilation operator that connects the zero - hole ( 4-electron ) and two - hole ( 2-electron ) groundstates of the @xmath12 plaquette must transform as @xmath13 . the same calculation can be performed for the hubbard model on a single plaquette ; while the wave - functions are somewhat more complex , in this case , the symmetry of the 2 and 4 electron ground - states are invariant for any @xmath1 in the range @xmath14 . as carlson et al . @xcite noted , it showed the robustness of the @xmath5-wave character of the pairing in @xmath2 and hubbard models . of course studies of a 4-site model could not say anything about the possibility of superconducting order . however , it turned out that studies of 2-leg @xmath2 and hubbard ladders yielded important insights concerning the character of the superconducting groundstate . the study of 2-leg ladders was motivated by a simple picture based upon the case in which the rung exchange interaction @xmath15 is large compared to the near neighbor leg exchange @xmath16 . in this limit , for the undoped heisenberg ladder , spin singlets tend to form on the rungs leading to a spin gapped groundstate . then when holes are added , where @xmath17 , they would occupy sites on either side of a rung so as to break only one exchange rung coupling . a measure of the spatial correlation of these rung hole pairs would then allow one to probe the superconducting order . based on this large @xmath18 picture , it was initially a surprise when numerical calculations @xcite showed that at half - filling the spin gap persisted to small values of @xmath19 . in addition , for the hole doped ladder , despite the fact that @xmath20 , the equal time pairfield - pairfield correlations appeared to have a power law decay , indicative of quasi - long - range superconductivity " . in later work @xcite , it was understood that the ladder would have a spin gap at half - filling for any finite @xmath21 , and that the groundstate of the doped ladder is a luther - emery @xcite liquid . furthermore , the pair structure is @xmath5-wave - like in the sense that the rung and leg pairfield amplitudes have opposite signs . we now also know that , in the limit as the length of the ladder tends to infinity , the @xmath2 ladder has perfect andreev reflection in response to an externally applied pairfield at one end of the ladder@xcite . the 2-leg @xmath2 and hubbard ladders now represent some of the best understood models of strongly correlated electron systems . admittedly , since the plaquette and the ladder are , respectively , zero and one dimensional systems , neither can support a superconducting phase with a finite transition temperature . however , in many cases it is possible to analyze the phase diagram of a higher dimensional system constructed as an array of _ weakly _ coupled mesoscale structures , starting from the exact numerical solution of the isolated structure , and treating the coupling between clusters in the context of perturbation theory @xcite . studies of arrays of weakly coupled two - leg ladders@xcite and plaquettes@xcite ( the `` checkerboard hubbard model '' ) lead to rather complex phase diagrams with many competing phases , even where the above analysis shows strong superconducting correlations on the isolated cluster . nonetheless , among those phases there are robust regions of @xmath5-wave , or @xmath5-wave - like superconductivity . as a final example of insights gained from studies of small systems , we turn to calculations on a 2-leg ladder model of an fe - pnictide superconductor @xcite . figure [ fig:1 ] shows the typical fermi surfaces of the fe - pnictide materials in an unfolded ( 1 fe / cell ) brillouin zone . there are two - hole fermi surfaces @xmath22 and @xmath23 around the @xmath24 point and two - electron fermi surfaces @xmath25 and @xmath26 around @xmath27 and @xmath28 . the symbols indicate the dominant @xmath5-orbital contributing to the bloch state on the indicated portion of the fermi surfaces . in weak coupling , rpa @xcite and functional renormalization group @xcite calculations suggest that the pairing arises from the scattering of time - reversed - pairs from the @xmath29-dominated states on the @xmath22 fermi surface to paired states with the same orbital character on the @xmath26 fermi surface , and from the analogous processes involving pairs in the @xmath30 dominated states on the @xmath22 and @xmath25 fermi surfaces . this is illustrated in fig . [ fig:1 ] for the @xmath31 pair scattering . ( solid line ) , @xmath30 ( dashed line ) and @xmath32 ( dotted line ) . the arrows illustrate the type of @xmath31 inter - fermi surface scattering processes that lead to pairing in the spin - fluctuation - exchange calculations.,scaledwidth=70.0% ] in order to use numerical methods to study these processes in the intermediate to strong coupling limit , the problem needs to be simplified . if we accept that the type of scattering processes shown in fig . [ fig:1 ] capture the essential physics , we can focus exclusively on pair scattering involving two bands and only one orbital . the resulting two - leg hubbard ladder retains the @xmath29 states along two cuts through the 2d bz , @xmath33 which passes through the @xmath22 fermi surface , and @xmath34 which passes through @xmath26 . this reduces the problem to that of the two - leg hubbard ladder shown in fig . [ fig:2]a which can then be studied using the numerical density matrix renormalization group ( dmrg ) @xcite . , @xmath35 and @xmath36 in units where @xmath37 . these hopping parameters were chosen to fit the dft calculation of the bandstructure for cuts with @xmath38 and @xmath39 . ; b ) the spin structure @xmath40 induced on the undoped fe ladder when an external magnetic field is applied to the lower left hand site . ; c ) the singlet pairfield @xmath41 induced across a rung , along a diagonal and along a leg at a distance 10 sites removed from the end of a @xmath42 fe - ladder with a unit external pairfield applied to its end rung.,scaledwidth=70.0% ] the one - electron hopping parameters @xmath43 , @xmath35 in units of @xmath44 , were chosen to reproduce the density functional bandstructure @xcite near the fermi surface for @xmath38 and @xmath39 . the repulsion @xmath1 between two electrons in the same orbital was varied in the range 34 in units of @xmath44 . in the undoped , one electron per site , limit one finds the expected spin gapped groundstate . by applying a magnetic field to one of the end sites of the ladder , the resulting expectation value of the spin appears as shown in fig . [ fig:2]b . here one sees stripe "- like @xmath28 spin correlations which decay with a slow exponential due to the spin gap . hole doping the system and applying an external pairfield on the end rung of the ladder , one obtains the pairfield singlet amplitudes illustrated in fig . [ fig:2]c . here a pairfield of unit strength was applied to the left end rung and fig . [ fig:2]c shows the strength of the induced pairfields @xmath41 ten sites to the right . the relative positive sign of the pairfield on the rung and diagonal and the negligible value of the pairfield on the legs is expected if the gap changes sign between the @xmath22 and @xmath26 fermi surfaces @xcite . the stripe - like sdw pattern of the spin correlations in the undoped system as well as the structure of the pairfield are consistent with what is found in the rpa calculations @xcite . however , what we found most interesting was the relationship between the fe - ladder and the previously studied 2-leg cuprate ladder . this is illustrated in fig . [ fig:3 ] . a has been twisted by @xmath45 and the phases of the orbitals denoted by the shaded sites have been changed by @xmath46 . ; b ) the spin expectation values of fig . [ fig:2]b for the twisted ladder show the spin gapped @xmath47 antiferromagnetic behavior of the familiar cuprate ladder . ; c ) the induced pairfield correlations of fig . [ fig:2]c become the familiar @xmath5-wave - like pairing correlations seen for models of the cuprate ladders.,scaledwidth=70.0% ] here in fig . [ fig:3]a , every other rung has been twisted by @xmath45 and the phase of the @xmath29-orbit has been changed by @xmath39 on each of the sites of the twisted rung . in this way , the rung one - electron hopping matrix element remains @xmath44 , but the leg and diagonal hoppings are changed to @xmath48 and @xmath49 , respectively . the dominant hoppings on the twisted fe - ladder are along the legs and rungs with only a weak diagonal hopping . the spin correlations shown in fig . [ fig:3]b , obtained by twisting every other rung of fig . [ fig:2]b , look just like the spin gapped af correlations of the previously studied 2-leg hubbard cuprate ladder . because of the twist and the phase change @xmath46 of the orbitals on the sites of the twisted rungs , the pairfield correlations take on the @xmath5-wave - like form shown in fig . [ fig:3]c . in short , the twist maps @xmath27 magnetic and sign - changing @xmath6-wave pairing correlations on the fe - ladder into @xmath47 magnetic and @xmath5-wave - like pairing correlations in the cuprate ladder ! finally , it turns out that the ratio of the leg - to - rung hopping 0.57 obtained from the fit to the fe - pnictide dft bandstructure is near the value which was previously found @xcite to give the slowest pairfield decay for a cuprate ladder . thus this fe - ladder turns out to simply be a twisted version of the cuprate 2-leg hubbard ladder with parameters near those which are optimal for pairing . this provides a direct link between the physics of these two materials . now , as noted by joe in his book _ introduction to mesoscopic physics _ @xcite , the interest in studying systems in the intermediate size range between microscopic and macroscopic is not only in order to understand the macroscopic limit . many novel phenomena exist that are intrinsic to mesoscopic systems . " here we have only touched on some examples where strongly correlated mesoscopic models have been introduced in the hope that they can provide some insight into the macroscopic high @xmath0 problem . it is natural to ask whether there are nt novel mesoscopic phenomena as well . indeed , there are . for example , the difference between the even- and odd - legged heisenberg ladders in which the even - leg ladders have a spin gap while the odd - leg ladders are gapless is a mesoscopic width effect @xcite . it is also known that while the doped 2-leg ladder goes into a luther - emery phase @xcite , it takes a finite doping to bring the 3-leg ladder into this phase @xcite . ladders also appear in the striped phase of the cuprates and , a better understanding of the mesoscopic properties of multi - leg ladders may shed light on the recently proposed @xmath39-phase shifted @xmath5-wave stripes @xcite .
the circadian clock is one of the most remarkable cyclic behaviors ubiquitous to the known forms of life , ranging from the unicellular to the multicellular level including prokaryotes . because of its importance , the underlying chemical reactions have been the subject academic interest for a long time and have recently been elucidated experimentally . circadian clocks have three important features : + 1 . they persist in the absence of external cues with an approximately 24-h period , which is rather long compared with most chemical reactions . they can be reset by exposure to external stimuli such as changes in illumination ( dark / light ) or temperature . the period of the circadian clock is robustly maintained across a range of physiological temperatures . ( temperature compensation ) @xcite . the emergence of cyclic behavior with a capacity for entrainment is theoretically understood as being result of the existence of a limit - cycle attractor in a class of dynamical systems described by chemical rate equations . however , the phenomenon of temperature compensation is not yet fully understood . generally , the rate of chemical reactions depends strongly on the temperature . most biochemical reactions , in particular , have an energy barrier that must be outcome with the aid of enzymes , and thus the rate can be expected to follow the arrhenius form . thus , the period of chemical or biochemical oscillators can be expected to strongly depend on the temperature @xcite . thus , the ubiquitous temperature - compensation ability of biological circadian clocks , suggests that there may be some common mechanism(s ) behind it . overall , there are two possibilities : one is that compensation exists at each elementary step , and the other is that compensation occurs at the system level for the total set of enzymatic reactions . recently , it was found that the rates of some elementary reaction steps in circadian clocks depend only slightly on temperature @xcite , suggesting that the activation energy barrier of some reactions is rather low . although some element - level compensation is important , it is difficult to imagine that every reaction step is fully temperature - compensated at the single - molecule level . indeed , if that were the case , temperature changes would not influence the oscillation of chemical concentrations at all . however , even though the period of oscillation is generally insensitive to temperature changes , it is known that the amplitude changes indeed depend on temperature @xcite . furthermore , circadian clocks are known to entrain to external temperature cycles , and they can be reset by temperature cues @xcite . thus , although temperature changes can influence the oscillation , the period is still robust . hence , it is necessary to search for a general logic that underlies the temperature compensation phenomenon at the system level . indeed , several models have been proposed @xcite . in most of the studies , several processes that are responsible for the period are considered , which cancel out the temperature dependence with each other . with such balance mechanism , the temperature compensation is achieved . these mechanisms , however , need fine - tuned set of parameters , or rather ad - hoc combination of processes for the cancellation . considering the ubiquity of temperature compensation , a generic and robust mechanism that dose not require tuning parameters is desirable . here we propose such a mechanism that has general validity for any biochemical oscillator consisting of several reaction processes catalyzed by enzymes . the mechanism can be briefly outlined as follows : biochemical clocks comprise multiple processes , such as phosphorylation and dephosphorylation , with each generally having a different activation energy barrier @xmath0 . the rate of such reactions , then , is proportional to @xmath1 , where @xmath2 as the concentration of `` free '' enzyme available . if the enzyme concentration were insensitive to temperature , the rate would just agree with the simple arrhenius form , thus implying high temperature dependence . now , consider a situation where several substrates share the same enzyme . at lower temperature , the reactions with higher activation energy will be slower and the substrates involved in these reactions will accumulate . then , because they share the same enzyme , competition for the enzyme will also increase . accordingly , the concentration of available ( free ) enzyme decreases , and when it reaches a level satisfying @xmath3 , these enzymatic reactions will be highly suppressed . the system spends most of its time under such conditions , which limits the rate . thus , the arrhenius - type temperature dependence is compensated for by the concentration of available enzyme as @xmath3 . although this is a rather basic description , its validity may suggest that temperature compensation emerges in any general chemical oscillation consisting of steps , catalyzed by a common enzyme . here , we first study the validity of this enzyme - limited temperature compensation mechanism for the specific case of kai protein clock model introduced by van zon et al . @xcite to explain the circadian clock of kaiabc proteins in cyanobacteria , which was discovered by kondo and his colleagues @xcite indeed , in this system , the period of circadian oscillation of kai proteins is temperature compensated @xcite . further , some of the elementary components in this system , specifically , kaic s atpase activity and kaic s phosphatase activity , were suggested to depend only slightly on temperature @xcite , but the origin of system - level temperature compensation has not yet been explained . here , we show numerically that system - level temperature compensation emerges from the differences in activation energy between phosphorylation and dephosphorylation and competition for kaia as the enzyme that catalyzes kaic phosphorylation . furthermore , we elucidate the conditions necessary for this temperature compensation to work . then , based on this analysis , we introduce a simpler model consisting of few catalytic reactions , to illustrate the above mechanism . possible relations between our results and reported experimental findings for circadian clocks are also discussed . the kai - protein - based circadian clock , discovered by kondo s group consists of kaia , kaib , kaic proteins with atp as an energy source @xcite . kaic has a hexameric structure with six monomers , each with two phosphorylation sites @xcite . it has both self - kinase activity and self - phosphatase activity , but the self - phosphatase activity is usually stronger , and so it is spontaneously dephosphorylated @xcite . kaia , in a dimer form @xcite , attaches to kaic and thus increases its kinase activity , leading to phosphorylation of kaic @xcite , while kaib inhibits the activity of kaia @xcite . this phosphorylation / dephosphorylation process of the kaiabc proteins constitutes a circadian rhythm . here , we simplify the process , to focus on the temperature compensation of the period . we reduce the two phosphorylation residues to just one , because abundance of singly phosphorylated kaic is strongly correlated with that of doubly phosphorylated kaic @xcite so that the phosphorylation of the two residues are equilibrated on a rather short time scale . next , we do not include kaib explicitly in our model , because changes in the concentration of kaib affect the period only slightly @xcite . note that although kaib is necessary to generate circadian oscillations , the effect can be accounted for by introducing a parameter value for kaia activity . here , we adopt a slightly simplified version of the model introduced by van zon , et al . @xcite(see also @xcite ) first , each kaic monomer has two states active and inactive . second , allosterically regulated kaic hexamers in the active state can be phosphorylated , whereas those in the inactive state can be dephosphorylated . a phosphorylated kaic monomer energetically prefers the inactive state , whereas a dephosphorylated kaic has the opposite tendency . here the flip - flop transition between active and inactive states occurs only from the fully phosphorylated or fully dephosphorylated states , as assumed in the concerted mwc model @xcite . no intermediate states are assumed . hence , the reaction process exhibits a cyclic structure as in fig . 1 . , while in the inactive state , dephosphorylation progresses without any enzymes at the rate @xmath4 . affinity between active kaic and kaia reduces as the number of phosphorylated sites of kaic increases successively . , title="fig : " ] next , kaia facilitates phosphorylation of active kaic with an affinity that depends on the number of phosphorylated residues of each kaic hexamer . kaics with a smaller phosphorylation number have stronger affinity to kaia and are phosphorylated faster . this assumption is necessary for generating stable oscillations @xcite . then , the reactions are given by @xmath5 @xmath6 @xmath7 @xmath8 { } } } ac_i \xrightarrow{k_p } c_{i+1 } + a\ ] ] here @xmath9 and @xmath10 denote the concentrations of active kaic and inactive kaic , respectively , with @xmath11 phosphorylated sites ; @xmath12 denotes the concentration of free kaia dimer . to study the temperature compensation of the period , we must also account for the temperature dependence of the reaction rate . here , the rates of phosphorylation and dephosphorylation are governed by the arrhenius equation . then the rate constants @xmath4 and @xmath13 are as follows : @xmath14 @xmath15 + with inverse temperature @xmath16 ( @xmath17 ) by taking the unit of boltzmann constant as unity . we could include the temperature dependence of the rates @xmath18 and @xmath19 in a similar manner , but as the reaction between active and inactive states progresses faster and does not influence the period @xcite , this dependence is neglected . van zon et al . demonstrated that temperature compensation occurs when the speed of phosphorylation and dephosphorylation is completely temperature - compensated at a level of elementary reaction process . however , it can not explain why the amplitude of oscillation depends on temperature , or the entrainment to temperature cycle occurs . the compensation mechanism at a system - level is wanted . here , the formation and dissociation of kaiac complexes occur at much faster rates than other reactions and so are eliminated adiabatically . thus , the change in the concentration @xmath12 is given by @xmath20 + where @xmath21 are the dissociation constants . considering the increase in affinity for kaia with the number of phosphorylated sites , we set @xmath22 . we adopted the deterministic rate equation given by the mass - action kinetics , and it is simulated by using the fourth - order runge - kutta method . as mentioned , a kaic allosteric model was analyzed . specifically , we study the case in which the activation energy for phosphorylation , @xmath23 , is larger than that of dephosphorylation , @xmath24 . see the supplementary information for other cases . ) . ( a ) red line indicates the time series of the mean phosphorylation level defined by @xmath25 , whereas the green line indicates that of the fraction of free kaia , @xmath26 . a decrease in temperature causes a decrease in the amplitude of the phosphorylation level . ( b ) time course of the abundance of each form of kaic , @xmath9 . at low temperature , the basal amount of @xmath27 ( magenta line ) is remarkably high . , title="fig : " ] for a certain range of parameter values , we found periodic oscillation in the kaic phosphorylation level and free kaia abundance , as shown in the time series in fig . the oscillation is described by a limit - cycle attractor , as represented in the orbit in a two - dimensional plane of kaic phosphorylation level and free kaia abundance . as the temperature increases , the amplitude of the limit - cycle increases ( fig . lowering the temperature causes a decrease in the maximum amount of free kaia and increase in the minimum level of kaic phosphorylation . further lowering it , however , results in the limit - cycle changing into a stable fixed point via hopf bifurcation . the timeseries of @xmath28 and the temperature dependence are shown in fig . 2b for @xmath16 = 1.0 , 1.5 , 2.0 . with a decrease in temperature , we see an increase in @xmath27 , that is , the abundance of kaic with five phosphorylated residues ( 5p - kaic ) , which leads to an increase in the minimum kaic phosophorylation level , as shown in fig . 2b . note that there is a remarkable change in the time course of @xmath27 ( the abundance of 5p - kaic ) at @xmath29 ( fig . 4a ) . below this inverse temperature , the minimum @xmath27 is close to zero . at higher @xmath16 ( i.e. , lower temperature ) , however , @xmath27 never comes close to zero , and the minimum value increases with lowering of temperature . . ( a ) maximum and minimum values of mean phosphorylation level over a cycle . the maximum value is nearly constant against temperature changes , whereas the minimum value increases with @xmath16 above @xmath30 , i.e. , at temperatures below the characteristic temperature @xmath31 . the oscillation disappears via hopf bifurcation at @xmath32 . ( b ) period of the oscillation . at high temperature ( low @xmath16 ) , the period changes exponentially with @xmath16 , whereas below @xmath31 , the period is nearly constant against changes in temperature . , title="fig : " ] ( a ) , @xmath33 ( b ) , @xmath26 ( c ) plotted against the inverse temperature @xmath16 . ( a)(b ) maximum ( red ) , minimum ( green ) , and average ( blue ) over a cycle . ( a ) note that the maximum value of @xmath27 is nearly constant , wheras its minimum and average increase with @xmath16 beyond @xmath34 ( i.e. , at lower temperature ) . ( b ) the average value of @xmath33 is fitted well by the value of the unstable fixed point of the equation ( magenta line ) , which is proportional to @xmath35 . ( c ) the concentration of free kaia in the plateau region of @xmath27 is plotted , as estimated from the time where @xmath36 reaches a peak . this free kaia concentration closely follows @xmath37 ( magenta line ) , for @xmath38 , i.e. , below the characteristic temperature . , title="fig : " ] the transition at @xmath39 is also reflected in the temperature dependence of the period . we plotted the period of oscillation as a function of ( inverse ) temperature , together with the maximum and minimum kaic phosphorylation levels ( see fig . 3 ) . above @xmath40 , the temperature dependence of the period follows @xmath41 , as can be naturally expected from a reaction process with a jump beyond the energy barrier . however , at lower temperature , the period is no longer prolonged exponentially and is nearly constant . thus , the temperature compensation of the circadian period appears at lower temperature . there is also a clear difference in the amplitude of oscillation below and above @xmath31 . at higher temperature ( without temperature compensation ) , the amplitude of oscillation is almost constant over a large interval of temperatures . however , at lower temperature ( in the temperature - compensated phase ) , the amplitude decreases with lowering of temperature ; eventually , the oscillation disappears via hopf bifurcation . this decrease in amplitude is due to the increase in the minimum value of the kaic phosphorylation level , caused by the increase in the minimum abundances of @xmath27 . the temperature dependences of the abundance of each kaic and free kaia also show distinct behaviors below and above @xmath31 ( see fig . 4 , supplementary fig . the average and minimum abundances of @xmath27 increase remarkably with a decrease in temperature below @xmath31 , whereas the maximum hardly changes with temperature ( fig . 4a ) . however , the amplitude of the oscillation of @xmath42 has a peak at around @xmath31 , and the minimum value increases as the temperature decreases . the minimum as well as the plateau value of free kaia increase when the temperature is decreased . the minimum inactive kaic abundance ( independent of the residue number ) and @xmath43 showed behaviors similar to those of @xmath42 , whereas the average followed @xmath44 dependence throughout the temperature range . considering that the total amount of all kaics is conserved and the difference in activation energy , this dependence itself is rather natural . thus , the transition in the oscillation and temperature compensation behavior at low temperature is the salient feature of the present system . next , we analyzed the conditions for @xmath23 and @xmath24 , the activation energies for phosphorylation and dephosphorylation , respectively , to elucidate the temperature compensation . supplementary fig . 3 shows a plot for the region where temperature compensation appears in the parameter space of @xmath23 and @xmath24 . from supplementary fig . 3 , we observe that the temperature compensation appears in the regime @xmath45 . the overall periodic behavior is determined mainly by @xmath46 , rather than the individual magnitudes . when @xmath47 , there still exists a transition to a phase with weaker temperature dependence of period on lowering the temperature , but this effect is not sufficient to produce the temperature compensation ( see also fig . + ( see supplementary informations for the case with @xmath48 . ) ) and the abundance of total kaia ( @xmath49 ) . red : temperature - compensated oscillation with @xmath50 ( @xmath51 : the period ) below the characteristic temperature . green : oscillation showing transition at the characteristic temperature . blue : oscillation without transition , with a simple increase with temperature of the arrhenius form . black : no oscillations at all . as @xmath52 increases from 0 to 1 , we successively see temperature compensation , transition , and a disappearance of oscillation . ( for the case with @xmath48 , see supplementary information ) . as @xmath49 increases , the width of the temperature compensation region decreases . ( b ) effect of kaia increase on the period . an increase in kaia led to narrowing of the range of the periodic solution , and also the range of temperatures at which temperature compensation occurs . , title="fig : " ] as already mentioned by von zon et al . , oscillation of kaic abundance requires that the amount of kaia is less than that of kaic @xcite . an increase in kaia abundance leads to decrease in the period , finally leading to no oscillations . the range of temperatures where the oscillations exist narrows as kaia abundance increases , and the oscillation disappears at higher temperatures . here , the transition temperature @xmath31 shifts to lower temperatures ( see fig . 5b ) . furthermore , the temperature compensation at @xmath53 is lost kaia abundance is increased . although the temperature dependence of the period is weaker at @xmath53 , the dependence @xmath54 still exists with @xmath55 smaller than @xmath23 . we plotted the range where temperature compensation occurs at @xmath53 in the two - dimensional plane with kaia abundances and @xmath56 . we see that low level of kaia abundance is necessary for temperature compensation . since kaic is a hexamer , we adopted six phosphorylation sites in our model . however , to understand the biological significance of this number of sites , we examine models with a reduced number of phosphorylation sites ( supplementary fig . we find that reducing the number of phosphorylation sites from hexamer to pentamer , and then to tetramer , narrows the temperature range where oscillations exist and temperature compensation occurs . for the tetramer , temperature compensation is not observed even at @xmath53 . ) are entrained by the temperature cycle between @xmath57 and 1.7 within about 10 cycles . , title="fig : " ] it has been experimentally it is suggested that kaic s phosphorylation cycle is entrained to external temperature cycle @xcite . to examine such entrainment , we cycle the temperature between @xmath57 and @xmath58 ( i.e. , within the region in the temperature compensation ) periodically in time with a period close to that of the kai system ( 27 h ) . within about 10 cycles , the phase of oscillation of the kaic system is entrained with that of temperature , independently of the initial phase of oscillation . thus , entrainment to temperature cycle is achieved ( see fig . here we discuss how temperature compensation is achieved . as presented in the results section , two stages are necessary : transition in the temperature dependence of the period , and complete temperature compensation of the period at lower temperature . as seen in the phase diagram ( fig . 5a ) , the former requires that @xmath23 is sufficiently larger than @xmath24 , which , as will be shown , means that the phosphorylation process is rate limited . for the latter , the abundance of kaia should be sufficiently small so that there is limited free kaia that can be used for phosphorylation ( i.e. , competition for free kaia ) . as will be shown below , the abundance of free kaia decreases as the temperature increases , which compensates for the increase in the rate constant of the reaction . phase transition when there is a difference between the energy barriers for phosphorylation and dephosphorylation , the temperature dependence of the rate of each processes is different . roughly speaking , the time scale for the phosphorylation process changes in proportion to @xmath59 , whereas the dephosphorylation has @xmath60 , where @xmath12 is the concentration of free kaia . thus , there exists a characteristic temperature at which the two rates are comparable ; the rate - limited reaction switches at this temperature , where the phase transition to temperature dependence occurs . thus , @xmath61 where @xmath12 , the concentration of free kaia , is estimated from the steady - state solution of our model ( see eq.([a ] ) ) to afford @xmath62 if @xmath63 and @xmath64 if @xmath65 . thus , a sufficient difference in activation energy is necessary for the phase transition - like behavior because the critical temperature will diverge as @xmath66 . if the difference is small , the critical temperature goes beyond the temperature for the onset of oscillation and the transition never occurs . moreover , if @xmath49 is too large or too small , the critical temperature is lower or higher than the range where the oscillation exists , and thus for both the cases , the phase transition disappears . these estimates agree with the phase diagram in fig . 5a . temperature compensation by self - adjustment of the kaia concentration when the temperature is lower than the transition temperature , the phosphorylation process takes more time , and @xmath27 is accumulated before dephosphorylation from @xmath43 progresses , as already discussed ( see fig.2b ) . the increase in the abundance of active kaic leads to competition for kaia , and thus a decrease in free kaia . if the total kaia abundance is limited , the system reaches a stage where phosphorylation almost stops . this leads to the plateau in the time course of @xmath27 , as observed in fig . this drastic slow down of the phosphorylation process occurs when @xmath12 is decreased to the level at @xmath67 . thus , during the plateau in @xmath27 , this approximate estimate gives , @xmath68 , so that the temperature dependence of the phosphorylation rate @xmath69 is compensated for by the decrease in @xmath12 . this plateau region is rate - limited in the circadian cycle , making the whole period independent of temperature . in more detail , this compensation is also estimated as follows . the abundances of the inactive forms of kaic decrease with an increase in @xmath27 . considering the differences in speed between phosphorylation and dephosphorylation , the total inactive kaic abundance ( at the fixed point ) is estimated as @xmath70 + which is consistent with fig . the flow from inactive kaic is thus estimated by @xmath71 this flow starts the phosphorylation processes from @xmath72 , but is slowed down at some residue number @xmath73 . in the present model , this slow down starts at @xmath74 due to the paucity of free kaia . following the above estimate of flow , the maximum @xmath75 is estimated to be proportional to @xmath76 . now from eq.([a ] ) and @xmath77 @xmath78 and @xmath79 @xmath80 at this time , the abundance of free kaia can be estimated approximately as @xmath81 when @xmath49 is small enough , the minimum free kaia abundance is smaller than @xmath82,thus @xmath83 . @xmath84 therefore , when @xmath49 is small enough and @xmath23 is sufficiently greater than @xmath24 , the time scale of the phosphorylation process is @xmath85 , and the period of the cycle is temperature - compensated . indeed , as shown in supplementary fig . 2 , the maximum @xmath86 ( and @xmath87 ) values show @xmath76 dependence , whereas free kaia shows approximately @xmath41 dependence when the phosphorylation is slowed down . in essence , the temperature compensation mechanism requires two properties : + difference in activation energy between phosphorylation and dephosphorylation processes , and a limited abundance of enzyme kaia . the former is essential to the phase transition , and the latter to the compensation of the arrhenius - type by the temperature dependence of the individual reactions . as long as these conditions are met , the period is temperature - compensated at low temperature , without the need for fine tuning of the parameters of the system . these two properties generally appear if there are two types of processes with different activation energy , with one type catalyzed by a common enzyme and the other without catalyzation . if the enzyme abundance is limited , the competition for the enzyme will lead to temperature compensation in cyclic reaction systems . in particular , if the main component of the chemical reactions has allosteric structure like kaic , the competition for enzymes will occur naturally . are catalyzed by the same enzyme @xmath88 , but the other reactions are not . the affinity between @xmath89 and @xmath88 weaken as @xmath11 increases . ( b ) maximum and minimum values of @xmath90 over a cycle and ( c ) the period , plotted against the inverse temperature @xmath16 . the oscillation is temperature - compensated over a specific range of temperatures , although the transition is not as sharp as in the kai model . , title="fig : " ] as an example , we consider the cyclic process shown in fig . 7 , where a cyclic change in five substrates occurs . three processes@xmath91have a high activation energy barrier , and are catalyzed by the common enzyme @xmath88 . in this case , the period of the oscillation in the concentration of each substrate is temperature - compensated at low temperature . ( if the number of reactions with the common enzyme is decreased to two , compensation still appears , but it is weaker , as shown in supplementary fig . 5 ) . this example demonstrates the generality of the present mechanism , and opens the possibility of applications to temperature compensation in other biochemical oscillations as well . the regulation of reaction process by autonomous changes in enzyme concentration , as adopted here , was previously pointed out by awazu and kaneko , who reported that relaxation to equilibrium slowes down when the concentrations of substrate and enzymes are negatively correlated @xcite . excess substrate hinders the enzymatic reaction , leading to a plateau in relaxation dynamics . in the present model , the total concentration of kaia as an enzyme is a conserved quantity but the fraction of free kaia available is reduced when there is abundant active kaic and so the speed of phosphorylation slows down dramatically . such regulation of reaction speed by limitation of catalysts should be important , not only for the temperature compensation of circadian clocks in general , but also for other biological processes @xcite . now we briefly discuss the relevance of the present results to experiments on the kai protein circadian system . in our model , it is assumed that kaic s affinity to kaia depends on the phosphorylation levels . this assumption is necessary not only for the emergence of temperature compensation but also for the existence of the oscillation itself . recent reports that kaic phosphorylation induces conformational changes may account for such changes in the binding affinity . in the actual circadian clock comprising kai proteins , kaib is also involved besides kaia and kaic . however , as mentioned , an increase in kaib abundance has minor influence on the clock @xcite . earlier theoretical study suggested that kaib binds to kaiac complexes strongly and restricts the concentration of free kaia @xcite . thus , the inclusion of kaib is expected to not alter the present temperature compensation mechanism , but to facilitate it by strengthening the limitation of enzyme availability . note that previously reported temperature compensation at element - level kaic s atpase activity @xcite and auto - dephosphorylation activity @xcite is also relevant to the system - level compensation in our model . the former is relevant to achieve fast equilibration between kaia and kaia complex used for adiabatic elimination in eq.([a ] ) , and the latter to provide @xmath92 . it is known that the amplitude of kaic oscillation decreases as the temperature is lowered @xcite , which agrees well with our results ( fig . 2a ) . indeed , as described already , this decrease in amplitude is tightly coupled with the temperature compensation mechanism . it is also interesting to note that in many circadian clocks , the period does not depend on the temperature , although the amplitude decreases with temperature @xcite . if every elementary step of the circadian clock were temperature compensated at the single - molecule level , this temperature dependence of the amplitude would not be possible furthermore , the entrainment of the kaic oscillation to imposed temperature cycles , as observed in a recent experiment would also not be possible @xcite . in our study , the compensation is based on the different temperature dependences of the phosphorylation and dephosphorylation processes ; the entrainment is a result of this difference . our mechanism for temperature compensation depends on paucity of kaia increasing the kaia concentration leads to loss of the compensation . we expect that this prediction will be directly confirmed in a future experiment . the authors would like to thank a. awazu , h. iwasaki , y. murayama , h. r. ueda , t. yomo for useful discussion . ( @xmath95 , @xmath96 ) . ( a ) maximum and minimum mean phosphorylation levels over a cycle . the maximum value is nearly constant against temperature changes . the minimum value increases with @xmath16 for @xmath97 , and is nearly constant below @xmath98 . the oscillation disappears at @xmath99 . ( b ) period of oscillation . the period changes with the arrhenius form , i.e. , in an exponential manner over the whole range of temperature , while the slope changes below and above @xmath31 . below @xmath31 ( beyond @xmath98 ) , the temperature dependence of the period is about @xmath100 , and is @xmath54 with @xmath55 larger than @xmath24 at @xmath101 ( @xmath102 ) . , title="fig : " ] and @xmath24 , while @xmath49 is fixed at @xmath105 . red : temperature - compensated oscillation satisfying @xmath106 ( @xmath51 : the period ) below the characteristic temperature . gray : oscillation without temperature compensation . when @xmath47 , the oscillation is temperature - compensated at temperatures below the characteristic temperature . , title="fig : " ] , while a pentamer and tetramer are the reduced models containing @xmath107 and @xmath108 , respectively . the same parameter values are used as in the original model . models with fewer phosphorylation sites ( @xmath109 ) can not generate any oscillation . in the reduced models , the temperature ranges where oscillations occur and where temperature - compensated oscillations occur are both narrower , and the temperature compensation in the tetramer model is not perfect . , title="fig : " ] are catalyzed by the same enzyme @xmath88 , and other reactions are not . the affinity between @xmath110 and @xmath88 is higher than that between @xmath111 and @xmath88 . ( b ) the period , plotted as a function of the inverse temperature @xmath16 . the range with temperature compensation is narrower , and the compensation is not perfect . ( inset : maximum and minimum values of @xmath111 over a cycle ) . , title="fig : " ]	circadian clocks ubiquitous in life forms ranging bacteria to multi - cellular organisms , often exhibit intrinsic temperature compensation ; the period of circadian oscillators is maintained constant over a range of physiological temperatures , despite the expected arrhenius form for the reaction coefficient . observations have shown that the amplitude of the oscillation depends on the temperature but the period does not this suggests that although not every reaction step is temperature independent , the total system comprising several reactions still exhibits compensation . we present a general mechanism for such temperature compensation . consider a system with multiple activation energy barriers for reactions , with a common enzyme shared across several reaction steps with a higher activation energy . these reaction steps rate - limit the cycle if the temperature is not high . if the total abundance of the enzyme is limited , the amount of free enzyme available to catalyze a specific reaction decreases as more substrates bind to common enzyme . we show that this change in free enzyme abundance compensate for the arrhenius - type temperature dependence of the reaction coefficient . taking the example of circadian clocks with cyanobacterial proteins kaiabc consisting of several phosphorylation sites , we show that this temperature compensation mechanisms is indeed valid . specifically , if the activation energy for phosphorylation is larger than that for dephosphorylation , competition for kaia shared among the phosphorylation reactions leads to temperature compensation . moreover , taking a simpler model , we demonstrate the generality of the proposed compensation mechanism , suggesting relevance not only to circadian clocks but to other ( bio)chemical oscillators as well .
high - altitude balloon experiments operating in stratosphere provide us with a unique opportunity to observe the sky in mostly any wavelength band with low airmass . as we are operating above 99% of the atmosphere , these experiments allow diffraction - limited operations without the use of adaptive optics . balloon payloads also experience lower downwelling radiation by a factor of more than ten as compared to the ground@xcite . floating zero - pressure balloons provide continuous observations varying from a few hours to days , depending on the launch site . the lower cost and the flexibility in launch timings and flight duration make high - altitude balloon experiments a wonderful test bed for space technologies@xcite . telescopic platforms at high altitudes have significant advantages over operations from the ground enabling observations at wider wavelength range . the weather conditions at these altitudes are also ideal , where the absence of clouds , water vapour and dust provide virtually ideal atmospheric transmission . the wavefront distortion at heights above 36 km is also very small . wavelengths below 400 nm are difficult to observe from ground - based observatories , but at balloon floating altitudes observations above 280 nm , as well as between 200 and 220 nm , become possible , the latter due the narrowing of o@xmath1 and o@xmath0 absorption bands@xcite . thus , a uv telescope ( 200400 nm ) located at these altitudes with aperture of just 6-inch in diameter , with sufficient pointing stability / accuracy and a 1k@xmath2k ccd array , could provide wide - field images with fwhm better than @xmath3 , approaching the diffraction limit@xcite ; similar to that of space observatories but at a much lower cost . additionally , space - based observatories have strict constrains because of detector safety reasons . these constrains on sun / moon angles limit observations of solar system objects . a stratospheric balloon payload typically has three following main components : a zero - pressure hydrogen / helium balloon of the required volume , a long flight train containing the recovery parachute and flight termination systems , and the gondola consisting of scientific equipment with communication and associated electronics@xcite . one of our major goals is to observe astronomical targets of opportunity from balloons . these objects appear with little warning and , as a result , are often only observed in the visible from ground - based observatories . with our quick response time we can have a balloon launch within two weeks , being well placed to obtain uv observations of these transients . a prime candidate for such observations would be comets which suddenly appear , as did comet ison ( c/2012 s1 ) or siding spring ( c/2013 a1 ) . these were pristine comets whose evolution as traced by the emission in the molecular bands of different ices should have revealed much about their composition . we have developed pointing / stabilization systems suitable for balloon flights which will allow us to observe comets and other transients ( e.g. supernovae ) in bands not observable by any other ground - based system . other high risk targets could be mercury / venus , whose observations are difficult because of the proximity to the sun . there are only a handful of uv observations of the moon ( surprisingly for such a prominent object in the sky ) , primarily because the moon is so bright that most spacecrafts have an avoidance zone around it . we propose to observe the moon both spectroscopically and with an imager to track changes in the albedo at different phases , where the phase angle of the solar illumination changes . there is an increased interest of the moon and it is important to understand the nature of the lunar surface@xcite . the complete payload . ] balloons can provide remote sensing and in - situ measurements of trace gases in upper troposphere and stratosphere@xcite . the uv window of 200400 nm includes the lines from several important species in atmospheric chemistry , such as s0@xmath0 , o@xmath1 , bro , hcho . spectroscopic observations at these wavelengths allow simultaneous measurement of multiple gas species , and thus avoid the need for separate electrochemical systems for each gas . balloon experiments are complementary to space - based instruments in that they are able to measure gas profiles with different observational geometries , while space - based instruments continuously monitor changes in the atmosphere over long periods of time . the widely accepted doas technique can be easily applied to spectroscopic observations of light in the atmosphere ( scattered or directly from the source ) in uv and visible wavelength regions to obtain the trace gas strengths@xcite . the overall structural payload design is shown in fig [ fig : payload ] . the payload will include several scientific instruments for simultaneous observations in both daytime and nighttime . the major instrument for the payload is a spherical catadioptric telescope of 80 mm aperture together with the uv spectrograph . the control of the pointing of the system is achieved through inertial measurement sensors ( coarse pointing ) and star sensor ( fine pointing ) . we will be operating this payload in two different modes . for bright sources , the telescope with a uv - sensitive ccd at the secondary focus will carry out imaging , while the spectrograph , with optical fiber connected to the light - collecting lens ( 20 mm aperture ) aligned with the telescope aperture , will conduct parallel spectroscopic observations of the same bright source . for faint sources observations , the optical fiber from the spectrograph will be placed at the secondary focus of the telescope enabling the spectroscopic observations . both modes can also be used for atmospheric observations . during day observations , the system will operate with only the coarse pointing as the star sensor will have difficulty identifying stars . individual components of the payload are described in detail below . the float altitudes have extreme environment conditions of low pressure , and the temperature can go down to @xmath4c or so . the low humidity and pressure at these altitudes demand extra care of high - voltage power supplies and detectors to prevent arcing . however , in contrast to space missions , the payload will experience lower vibrations and shocks . the system consists of an 80 mm spherical catadioptric telescope mounted on the pointing platform@xcite . it has a rectangular field of view of @xmath5 . the optical design is a two spherical mirror configuration and a double - pass corrector lens system . the secondary focus can house either a uv ccd with a suitable filter or a slit , depending on the mode of observation . the ccd is a @xmath6 pixels detector which operates at 12 frames per second ( fps ) . the instrument is designed to withstand loads up to 25 g. one of the main objectives is to scan the sky looking for transients . the instrument will acquire the images at a fast frame rate and will analyze each frame looking for brightness variations over the frames . the transient events will be stored on - board or transmitted back to the ground station through a radio link . .spectrograph technical specifications [ cols= " < , < " , ] we plan to fly the payload on a zero - pressure balloon capable of floating at @xmath7 km altitude for up to 5 hours with the national balloon facility station of tifr at hyderabad@xcite in november of 2016 . the duration of the float will enable us to observe in the night ( with the launch 34 hours before sunrise ) for astronomical objects , and at daylight for atmospheric studies . the reference spectrum for trace gas analysis will be obtained at float , and the measurements will continue during descent@xcite . in this flight we also plan to supplement the observational instruments with the detachable pre - sterilized sampling chamber _ sample _ ( stratospheric altitude microbiology probe for life existence)@xcite designed to collect and contain the dust stratospheric samples and get them back without contamination . november is the peak time for the leonid meteor shower , and we plan to collect cometary dust particles , where further study will be conducted to establish the possibility of microbial life in the upper atmosphere . we have presented an overview of our ballooning program and the instruments that currently are developed for astronomical observations . these instruments can also be used for atmospheric observations such as measuring the limb scattered uv radiance and trace gas strengths . in our earlier high - altitude flights , the payloads experienced frequent oscillations preventing us from continuous observation for a long duration . we have developed a pointing and stabilization platform with two - fold coarse / fine pointing , which will enable stable observations of regions of interest . we are testing and verifying the operation of our payloads on balloon flights with the intention of placing them on small satellites , such as , e.g. cubesats . going into space has the obvious advantages of longer mission life and the ability to go further into the far - uv . there is an increasing number of opportunities for small missions although without the quick response times possible with balloons , and the development of instrumentation skills will benefit our future access to space . hibbitts , c. a. , young , e. , kremic , t. and landis , r. , science measurements and instruments for a planetary science stratospheric balloon platform " , _ aerospace conference , ieee , big sky , mt _ , pp . 1 - 9 ( 2013 ) . safonova , m. , nayak , a. , sreejith , a. g. , mathew , j. , sarpotdar , m. , ambily , s. , nirmal , k. , talnikar , s. , hadigal , s , prakash . a. and murthy , j. , an overview of high - altitude balloon experiments at the indian institute of astrophysics " , _ astron . & astroph . _ , * 29 * , * in press * , ( 2016 ) . sreejith , a. g. , mathew , j. , sarpotdar , m. , nirmal , k. , suresh , a. , prakash , a. , safonova , m. , and murthy , j. , measurement of limb radiance and trace gases in uv over tropical region by balloon - borne instruments flight validation and initial results " , _ atmos . _ , * in review * , ( 2016 ) . platt , u. and stutz , j. : _ differential optical absorption spectroscopy : principles and applications , physics of earth and space environments _ , isbn 978 - 3 - 540 - 21193 - 8 , springer - verlag berlin heidelberg , ( 2008 ) . mathew , j. , prakash , a. , sarpotdar , m. , sreejith , a. g. , safonova , m. , and murthy , j , ultraviolet cosmic imager to study bright uv sources " , _ paper 9905 - 146 _ , space telescopes and instrumentation 2016 : ultraviolet to gamma ray , part of spie astronomical telescopes + instrumentation , ( 2016 ) . ambily , s. , sarpotdar , m. , mathew , j. , sreejith , a. g. , nirmal , k. , prakash , a. , safonova , m. , and murthy , j. , development of data acquisition methods for an fpga - based photon counting detector " , _ journal of astronomical instrumentation _ , * submitted * , ( 2016 ) . nirmal , k. , sreejith , a. g. , mathew , j. , sarpotdar , m. , suresh , a. , safonova , m. , and murthy , j. , pointing system for the balloon - borne telescope " , _ journal of astronomical telescopes , instruments , and systems _ , * submitted * , ( 2016 ) . sarpotdar , m. , mathew , j. , sreejith , a. g. , murthy , j. and safonova , m. , design and development of a star sensor cum asteroid tracker " , _ paper 9904 - 195 _ , space telescopes and instrumentation 2016 : optical , infrared , and millimeter wave , spie astronomical telescopes + instrumentation , ( 2016 ) . prakash a. , safonova m. , murthy j. , kumble s. , mathew j. , sreejith a.g . , sarpotdar m. , nirmal k. , ambily s. , chakravortty d.and rangarajan a. , collection of stratospheric samples using balloon - borne payload system " , 41st cospar scientific assembly , _ abstract i d # 19527 _ , 30 july7 august , istanbul , turkey , ( 2016 ) .	the ultraviolet ( uv ) window has been largely unexplored through balloons for astronomy . we discuss here the development of a compact near - uv spectrograph with fiber optics input for balloon flights . it is a modified czerny - turner system built using off - the - shelf components . the system is portable and scalable to different telescopes . the use of reflecting optics reduces the transmission loss in the uv . it employs an image - intensified cmos sensor , operating in photon counting mode , as the detector of choice . a lightweight pointing system developed for stable pointing to observe astronomical sources is also discussed , together with the methods to improve its accuracy , e.g. using the in - house build star sensor and others . our primary scientific objectives include the observation of bright solar system objects such as visible to eye comets , moon and planets . studies of planets can give us valuable information about the planetary aurorae , helping to model and compare atmospheres of other planets and the earth . the other major objective is to look at the diffuse uv atmospheric emission features ( airglow lines ) , and at column densities of trace gases . this uv window includes several lines important to atmospheric chemistry , e.g. so@xmath0 , o@xmath1 , hcho , bro . the spectrograph enables simultaneous measurement of various trace gases , as well as provides better accuracy at higher altitudes compared to electromechanical trace gas measurement sondes . these lines contaminate most astronomical observations but are poorly characterized . other objectives may include sprites in the atmosphere and meteor flashes from high altitude burn - outs . our recent experiments and observations with high - altitude balloons are discussed .
there are two dimensionless constants , the fine - structure constant ( @xmath0 ) and the proton - electron mass ratio ( @xmath2 ) , for which spectroscopy is a test ground to probe temporal and spatial variations . the constant @xmath0 is important for electronic structure of atoms and molecules . the dependence of the spectrum on @xmath0 appears through the relativistic corrections , such as the fine - structure , lamb shift , etc . the leading relativistic corrections scale as @xmath3 where @xmath4 is atomic number . therefore , heavier elements have higher sensitivity to @xmath0-variation . for the astrophysically relevant elements @xmath5 and relativistic corrections do not exceed few percent . the constant @xmath2 enters atomic spectra only through the isotope shifts which are typically on the order of @xmath6 , or less . due to the presence of the vibrational and rotational structures , the dependence of the molecular spectra on @xmath2 is much more pronounced . therefore , the search for @xmath0-variation is usually done using atomic spectra @xcite , while @xmath2-variation is studied with the help of molecular spectra @xcite . let us note in passing that much higher sensitivity to the variation of both fundamental constants than in optics , can be found in the infrared and microwave wavebands @xcite . astronomical differential measurements of the constant @xmath0 is based on the comparison of the line centers in the absorption or emission spectra of cosmic objects and the corresponding laboratory values . it follows that the uncertainties of the laboratory rest frequencies and of the line centers in astronomical spectra are the prime concern of such measurements . for example , the unknown isotopic abundances of the elements in the universe can lead to the systematic frequency shifts , comparable to the expected signal from @xmath0-variation @xcite . that is why it is important to use such heavy elements as fe and ni where relativistic effects are larger , while isotopic shifts are suppressed . most studies of the possible @xmath0-variation at the redshifts @xmath7 are based on the analysis of the quasar absorbtion spectra using many multiplet method suggested in refs . this method utilizes lines of different ions from the same source . all these lines are analyzed simultaneously to find the redshift and the value of @xmath0 for this object . the lines with small sensitivity coefficients to @xmath0-variation serve as anchors which give the redshift , while the lines with high sensitivity serve as probes , which give the value of @xmath0 . all lines of the light ions , like mg ii and al ii , belong to the first category . most lines of heavier ions , like fe ii , ni ii , and zn ii , fall into second category . however , some of the lines of ni ii have relatively small sensitivity and belong to anchors . this can be important for the control of the systematics . the lines of the ion fe ii have sensitivity coefficients to @xmath0-variation of different signs . this also allows effective control of the possible systematic effects . at present there is one group , that reports nonzero space - time variation of @xmath0 on the level of few parts per million ( ppm ) @xcite ( known as the `` australian dipole '' ) . these results have not been confirmed by other groups , who give only upper limits on the @xmath0-variation also at the ppm level ( see @xcite and references therein ) . because of that it is important to continue observations and include more sources and additional lines . such programs are currently going on at the vlt and keck telescopes @xcite . a list of the astrophysically relevant optical lines of atoms and ions is given in compilations @xcite . these compilations includes eleven lines of ni ii . according to @xcite there are nine lines of ni ii , which can be observed in the quasar spectra including one line not listed in @xcite . the rest frequencies for ni ii are tabulated in nist database @xcite . compilation by @xcite includes both frequencies and oscillator strengths @xmath8 for astrophysically relevant lines . several high accuracy laboratory rest frequencies were measured in ref.@xcite . according to @xcite more accurate laboratory measurements are needed . the sensitivity to @xmath0-variation ( @xmath1-factors ) was theoretically studied in refs . @xcite for four lines from the list @xcite . in this paper we do calculations of the @xmath1-factors and oscillator strengths @xmath8 for all nine astrophysically interesting lines of ni ii . supposing that the nowadays value of @xmath0 differs from its value in the earlier universe we can study space - time variation of @xmath0 by comparing atomic frequencies for distant objects in the universe with their laboratory values . in practice , we need to find relativistic frequency shifts , known as @xmath1-factors @xcite . the difference between the transition frequencies in astrophysical spectra and the laboratory ones is given by the formula : @xmath9 where @xmath10 is a transition frequency in non - relativistic approximation , @xmath11 is a relativistic correction . in eq . ( [ eq : theory1 ] ) @xmath12 is the frequency value for @xmath13 , where @xmath14 is fine - structure constant in the laboratory ( present time ) conditions . note that transition @xmath1-factor is simply a difference between @xmath1-factors of upper and lower levels . for the search of the @xmath0-variation it is most advantageous to use atoms and ions for which @xmath1-factors of transitions between certain states significantly differ from each other . that means we need elements with high @xmath4 . at the same time these elements should be abundant in the universe to provide sufficient observational data . the latter restriction leaves us with fe and ni as the most heavy abundant elements . fe has been studied in detail in refs . @xcite and other relevant elements have been studied in @xcite . ni has been investigated in a lesser detail @xcite and not all levels observed in astrophysics have been calculated . this is why we return to this problem here . rough estimates of the @xmath1-factors can be obtained from a simple one - particle model @xcite . but in order to obtain more accurate values one has to account for electronic correlations and perform large - scale numerical calculations . to find @xmath1-factors numerically we need to solve the atomic relativistic eigenvalue problem for different values of @xmath0 , or equivalently for different values of @xmath15 from eq . ( [ eq : theory1 ] ) . in this case we can get @xmath1-factor as : @xmath16 our previous experience shows that convenient choice is @xmath17 . in order to test the accuracy of this approximation we can also estimate the second derivative : @xmath18 where @xmath19 . when second derivative is small , results of the calculation using eq . ( [ eq : theory2 ] ) are sufficiently reliable . if this were not the case we would have strong interaction between levels . such situation requires more carefulness . it may be useful to trace the levels to smaller values of @xmath15 . therefore , we do additional calculation for @xmath20 . when theoretical splitting for the interacting levels at @xmath21 is close to the experimental one , then we can expect good accuracy for the @xmath1-factors even for this case . otherwise , if the splitting at @xmath21 differs from the experiment , the @xmath1-factors calculated for @xmath21 may be incorrect . we can try to improve calculated @xmath1-factors by moving along @xmath15 axis to the point where theoretical splitting matches experiment @xcite . as an additional test of the accuracy of our theory we compare calculated @xmath22-factors with the experimental ones from @xcite . absorption spectra in astrophysics correspond to the transitions from the ground state , which in the case of ni ii has @xmath23 and belongs to configuration [ ar]@xmath24 . since ni ii has nine electrons in the open shells its spectrum is dense and complicated @xcite . due to the proximity of the levels with the same total angular momentum @xmath25 and parity @xmath26 they may strongly interact with each other , in particular for the high frequencies which are more interesting for astrophysics . because of the selection rules for the transitions from the ground state we are interested in the states of the negative parity with @xmath27 . in the frequency range observed in the high redshift astrophysics ( i.e.roughly between 52000 and 70000 @xmath28 ) @xcite there are five lowest odd levels with @xmath29 , eight levels with @xmath23 , and seven levels with @xmath30 . all these levels belong to configurations @xmath31 , @xmath32 , and @xmath33 . we performed calculations for all these levels , as well as for the ground state for four values of @xmath15 and found transition frequencies . then we used cubic interpolation for the interval @xmath34 . we use dirac - coulomb hamiltonian in the no - pair approximation . breit corrections to the @xmath1-factors were studied previously and were found to be small @xcite . we apply configuration interaction ( ci ) method in the final basis set of relativistic orbitals . all calculations are done with the computer package of i. i. tupitsyn @xcite . we start with solving the hartree - fock - dirac equations . the self - consistency procedure is done for the ground state configuration [ ar]@xmath35 . after that all these shells are frozen . the valence orbital @xmath36 is constructed for the @xmath37 configuration . then two @xmath38 orbitals are found for the non - relativistic configuration @xmath31 . on the next stage we make virtual orbitals using the method described in @xcite . in this method an upper component of the virtual orbitals is formed from the previous orbital of the same symmetry by multiplication with some smooth function of radial variable @xmath39 in the spherical cavity with a radius 50 a.u . the lower component is then formed using kinetic balance condition . this way we make finite basis set of 33 orbitals that include @xmath40 partial waves ( @xmath41 ) . the configuration space is formed by making single and double ( sd ) excitations from the small list of reference configurations . for even states this list includes @xmath24 and @xmath37 . for odd states the reference configurations are : @xmath31 , @xmath32 , and @xmath33 . this way we get 3292 even and 2991 odd relativistic configurations . the number of configurations accounted for in our present calculations is significantly bigger than in the previous calculations in ref . @xcite . lcldrddrrrr & & & & + & & & & & & & & & & + @xmath42 & & 51558 & 1.420 & @xmath4349002 & 1.423 & 1.429 & @xmath44 @xmath45 & @xmath46 @xmath45 & @xmath47 & @xmath48 + @xmath49 & & 52739 & 1.365 & 50239 & 1.359 & 1.371 & @xmath50 @xmath45 & @xmath51 @xmath45 & @xmath52 & @xmath53 + @xmath54 & & 53635 & 1.186 & 51183 & 1.187 & 1.200 & @xmath55 @xmath45 & @xmath56 @xmath45 & @xmath57 & + @xmath58 & & 54263 & 1.025 & 51693 & 1.010 & 0.984 & @xmath59 @xmath45 & @xmath60 @xmath45 & @xmath61 & @xmath62 + @xmath63 & & 55019 & 0.616 & 52482 & 0.609 & 0.571 & @xmath64 @xmath45 & @xmath65 @xmath45 & @xmath66 & @xmath60 + @xmath67 & & 55418 & 1.184 & 53008 & 1.194 & 1.238 & @xmath68 @xmath45 & @xmath69 @xmath45 & @xmath70 & @xmath71 + @xmath72 & & 56075 & 0.985 & 53728 & 0.996 & 1.029 & @xmath73 @xmath45 & @xmath74 @xmath45 & @xmath75 & @xmath76 + @xmath77 & a & 56371 * & 0.940 & 53972 & 0.923 & 0.889 & @xmath78 @xmath45 & @xmath79 @xmath45 & @xmath80 & @xmath81 + @xmath82 & & 56425 & 0.412 & 54140 & 0.420 & 0.400 & @xmath83 @xmath45 & @xmath84 @xmath45 & @xmath85 & + @xmath86 & b & 57081 * & 1.154 & 54817 & 1.134 & 1.143 & @xmath87 @xmath45&@xmath88 @xmath45 & @xmath89 & @xmath90 + @xmath91 & c & 57420 * & 1.116 & 55315 & 1.100 & 1.200 & @xmath92 @xmath45 & @xmath93 @xmath45 & @xmath94 & @xmath95 + @xmath96 & d & 58493 * & 0.946 & 56376 & 0.966 & 0.857 & @xmath97 @xmath45 & @xmath98 @xmath45 & @xmath99 & @xmath100 + @xmath101 & & 58706 * & 0.795 & 56770 & 0.799 & 0.800 & @xmath102 @xmath45 & @xmath103 @xmath45 & @xmath104 & + @xmath105 & & 66571 * & 1.480 & 66169 & 1.506 & 1.600 & @xmath106 @xmath45 & @xmath107 @xmath45 & @xmath108 & + @xmath109 & e & 66580 & 1.550 & 66173 & 1.592 & 1.733 & @xmath110 @xmath45 & @xmath84 @xmath45 & @xmath111 & + @xmath96 & & 67695 & 0.960 & 67512 & 0.943 & 1.029 & @xmath112 @xmath45 & @xmath113 @xmath45 & @xmath114 & + @xmath86 & g & 68131 * & 1.200 & 67921 & 1.186 & 1.143 & @xmath115 @xmath45 & @xmath116 @xmath45 & @xmath117 & + @xmath101 & f & 68154 * & 1.020 & 68080 & 1.033 & 0.800 & @xmath118 @xmath45 & @xmath119 @xmath45 & @xmath120 & + @xmath91 & & 68736 * & 1.264 & 68753 & 1.242 & 1.200 & @xmath121 @xmath45 & @xmath122 @xmath45 & @xmath75 & + @xmath42 & h & 70778 & 1.385 & 70704 & 1.383 & 1.429 & @xmath123 @xmath45 & @xmath124 @xmath45 & @xmath125 & + results of our calculations of the energies , @xmath22-factors , @xmath1-factors , and their derivatives @xmath126 for the odd levels are presented in table [ tbl : q - factors ] and fig . [ alpha_dep ] . calculations were done for four values of @xmath15 : @xmath127 . then we did cubic interpolation for the interval @xmath128 . the @xmath1-factors were found by using eq . ( [ eq : theory2 ] ) and by differentiation of the interpolation polynomial at @xmath21 . both results appeared to be very close . identification of the levels is done by comparison of the calculated @xmath22-factors with the experimental ones and with the prediction of the @xmath129-coupling scheme ( see table [ tbl : q - factors ] ) . for the energy range in question the order of the calculated levels agrees with the experiment . moreover , one can see from the figure that our calculation in general reproduces the splittings between the nearest levels , but somewhat underestimates the energies of the lower levels of negative parity . for most levels the derivative @xmath130 is rather small and negative . that means that interaction between neighboring levels with the same quantum numbers @xmath131 is not very strong . however there are four levels in table [ tbl : q - factors ] with @xmath132 . in each of these cases there is close level with large negative @xmath130 . therefore , we can assume that these levels are at pseudo crossing and there is strong repulsion between them . such interacting pairs of levels are marked on fig . [ alpha_dep ] . below we analyze each of these pairs separately . [ [ levels - a2g_72-and - b2f_72 . ] ] levels a@xmath133 and b@xmath134 . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + this is the strongest interacting pair of levels with @xmath135=@xmath136@xmath28 and @xmath137=@xmath138@xmath28 . calculated energy splitting , 845 @xmath28 , is larger than experimental splitting 709 @xmath28 . it is clear from fig . [ alpha_dep ] that pseudo crossing takes place at positive values of @xmath15 . both calculated @xmath22-factors differ from the experiment by roughly 2% . if we move along @xmath15 axis towards the crossing point at @xmath139 we get smaller splitting and can expect better agreement with the experiment . indeed , the splitting for @xmath140 is 758 @xmath28 , which is almost three times closer to the experimental value . the error for @xmath22-factors also decreases to 1% , or so . we conclude that this way we do get better agreement with the experiment . at the pseudo crossing @xmath141 the levels are parallel to each other , so @xmath142 . therefore , the shift towards @xmath143 strongly affects calculated @xmath1-factors of each level leaving their sum almost constant . at @xmath140 we get @xmath144 @xmath28 and @xmath145 @xmath28 . to get exactly experimental splitting we need to move even closer to @xmath143 . however , the interaction of the levels a and b with other levels is not at all negligible and we need to reproduce other energy splittings as well . this can not be done with the help of the single parameter @xmath15 . therefore , we take @xmath1 factors at @xmath140 as our recommended values . theoretical error here is , of course , quite large . we estimate it to be around 300 @xmath28 , which covers the whole range of @xmath15 between @xmath21 and @xmath141 . [ [ levels - c2d_52-and - d2f_52 . ] ] levels c@xmath146 and d@xmath147 . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + these levels interact much weaker than the first pair with @xmath148=@xmath149@xmath28 and @xmath150=@xmath151@xmath28 . besides , the calculated energy splitting , 1061 @xmath28 , is very close to the experimental splitting 1073 @xmath28 . one of the two @xmath22-factors is 2% larger than experimental value , but another one is much closer . we conclude that no correction is necessary for this pair of levels . [ [ levels - e4p_32-and - f2d_32 . ] ] levels e@xmath152 and f@xmath153 . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + though these levels have derivatives of the opposite sign , their absolute values differ by a factor of 3 . that means that the two level model is not applicable here . the energy splitting is overestimated by 332 @xmath28 , or about 20% and @xmath22-factor of the level e is 3% larger than experimental value . because of the strong interaction of this pair with other levels we do not introduce any correction , but rather use the shift in @xmath15 to estimate the error . experimental splitting is reproduced at @xmath154 where @xmath1-factors appears to be : @xmath155 @xmath28 and @xmath156 @xmath28 . therefore , we estimate the error to be about 250 @xmath28 for both levels . [ [ levels - g2f_72-and - h4d_72 . ] ] levels g@xmath134 and h@xmath157 . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + interaction of these two levels is a little stronger than for the pair ( c , d ) and theoretical energy splitting , 2647 @xmath28 is 135 @xmath28 smaller , than required . calculated @xmath22-factors are quite good , but the difference between them is small and we can not use them to estimate the mixing of these levels with each other . the optimal splitting corresponds to @xmath158 where @xmath159 @xmath28 and @xmath160 @xmath28 . we estimate the error for the @xmath1-factors to be about 200 @xmath28 . for the remaining levels , which do not form strongly interacting pairs , the slope remains relatively constant . calculated @xmath1-factors are , therefore , less sensitive to the details of the calculation . we checked that any shifts along @xmath15 axis in order to improve energy splittings between neighboring levels do not change @xmath1-factors by more than 100 @xmath28 . thus , we estimate theoretical error for these @xmath1-factors to be around 150 @xmath28 . our final recommended values for the @xmath1-factors with the error bars are listed in table [ tbl : q - factors ] . approximately half of the levels from this table were studied before in ref.@xcite . for all of them we have good agreement between two calculations . at the same time both calculations do not agree with the earlier calculation @xcite . . oscillator strengths @xmath8 for the transitions considered in this paper . calculations were done in the length ( l ) and velocity ( v ) gauges . were possible we compare our results with compilation of @xcite and with ref . [ cols="<,<,^,^,^,^ " , ] [ [ oscillator - strengths . ] ] oscillator strengths . + + + + + + + + + + + + + + + + + + + + + observability of the transitions considered here depends on the respective oscillator strengths @xmath8 . not all of them are known from the experiment . we calculated e1 transition amplitudes and @xmath8 in the length ( l ) and velocity ( v ) gauges for all transitions from table [ tbl : q - factors ] . results are summarized in table [ tbl : fosc ] . for the transitions with @xmath161 two calculations agree within 10% . even for most weaker transitions @xmath162 the agreement is quite good . only in one case the difference is close to 30% . usually , for the ci wave functions the difference between amplitudes in l- and v - gauges can be used to estimate the accuracy of the theory . therefore we expect our results for the transitions with @xmath163 to be accurate within 20 - 30% . within this error bar our results agree with compilation of experimental and observational results by @xcite . the only exception is the line 58706 @xmath28 , where our strength is 50% larger than morton s . we also have mostly good agreement with ref . @xcite . for the line 68736 @xmath28 our values lie between those of refs . @xcite and @xcite . however , for the lines 55418 and 56371 @xmath28 our results are significantly smaller . in particular , for the potentially interesting line 56371 @xmath28 our strength is three times smaller than in @xcite . this makes observation of this line for the high redshift sources more difficult . to summarize , here we present theoretical @xmath1-factors for ni ii found with ci method for dirac - coulomb hamiltonian in no - pair approximation . we calculated @xmath1-factors for several new lines , which had not been studied theoretically before , but were observed in the high redshift quasar spectra . all calculated sensitivities for ni ii are negative . two new lines have relatively small @xmath1-factors , around @xmath164 @xmath28 and one has @xmath1-factor , which is one of the largest in absolute value , @xmath165 @xmath28 . note , that large difference in sensitivities of individual lines increases sensitivity of the observations to @xmath0-variation and allows effective control of the systematics . in comparison to the previous calculations we have significantly increased the ci space . that did not changed results very much and we have good agreement with the most resent calculation @xcite . we do not think the accuracy can be noticeably improved within ci method : ni ii has 9 electrons in open shells and it is practically impossible to saturate ci space . on the other hand , already present accuracy is sufficient to analyze astrophysical data on the possible @xmath0-variation . we thank h. rahmani and r. srianand for stimulating this research and j. berengut and m. murphy for interesting discussions . this work is partly supported by the russian foundation for basic research grant no.14 - 02 - 00241 .	we calculated the dependence of the transition frequencies on the fine - structure constant @xmath0 ( @xmath1-factors ) for ni ii . nickel is one of the few elements with high sensitivity to @xmath0-variation , whose lines are observed at high redshifts . this makes it a sensitive probe for @xmath0-variation on the cosmological timescale . the electronic structure of ni ii ion was treated within the configuration interaction ( ci ) method using dirac - coulomb hamiltonian .
feature finding algorithms are designed to automatically localize the intensity centroid of a specific feature in an image . in biological systems , the features of interest are often fluorescently labelled proteins and organelles , with dimensions below the diffraction limit ; the intensity distributions to be localized are therefore diffraction limited spots . feature finding algorithms consist of two main steps to automatically detect and then localize all features of interest in a given image . the first step is an initial detection of local intensity maxima over the entire image , identifying candidate feature locations to pixel resolution . in the second step , the full intensity distribution of the candidate features , which spans many pixels , is used to localize the centroid of each intensity distribution to sub - pixel resolution . to extract the real space dynamics of a feature from a time - series of images , tracking algorithms establish the correspondence between features in successive image frames , forming a set of single - particle trajectories . in general , feature correspondence is determined by minimizing the displacement of localized features between frames . the ability of tracking algorithms to reconstruct feature trajectories with high fidelity is critically dependent on the ability of the localization algorithm to determine the particle positions accurately . among articles that assess the limits of accuracy and precision of particle localization algorithms , cheesum _ et al . _ @xcite quantitatively compare algorithms for localizing fluorescently labelled objects in two - dimensions , from among centre - of - mass algorithms , direct gaussian fitting which exploits the shape of the intensity distribution for a diffraction - limited spot , image cross - correlation algorithms , and sum - absolute difference methods . typically , centre - of - mass and direct gaussian fitting methods obtain sub - pixel resolution directly . in contrast , both the cross - correlation and sum - absolute difference methods determine the object s position to only pixel resolution . to obtain sub - pixel resolution , the cross - correlation or sum - absolute difference matrices must be interpolated . measured accuracy as the mean difference between the localized position and actual position of a simulated object over a large number of trials and determined , unsurprisingly , the direct gaussian to be the superior algorithm for localizing point sources . _ examine the factors that limit the precision of centroid localization @xcite , for limiting cases where a large number @xmath1 of signal photons are collected , in which the data is considered photon shot - noise limited ; and the limit of small @xmath1 , where the image is considered background noise limited , obtaining the theoretical limiting precision for one spatial dimension , @xmath2 by considering each of these limits independently and linear interpolation between them @xcite . in a more general treatment , ober _ et al . _ have used a fisher information matrix to determine the limit of localization precision for single - molecule microscopy @xcite . the inverse of the fisher information matrix is independent of the particular localization procedure . however , importantly , ober _ et al . _ assumed the localization procedure would provide an _ unbiased _ estimate of the object location . this assumption is not justified for all localization routines , as was shown in @xcite and @xcite . the ultimate limit of localization precision , as a function of the emission wavelength @xmath3 , photon emission rate @xmath4 , objective lens numerical aperture @xmath5 , and acquisition time @xmath6 , is @xmath7 , where @xmath8 is the optical system efficiency , defined as the fraction of photons leaving the object that reach the detector . this is a fundamental limit of localization and does not include the effects of pixelation or noise @xcite . when these effects are included , the relation is much more complicated , and the limiting precision depends on the particular imaging conditions and detector specifications . knowledge of the expected intensity distribution for a diffraction - limited spot is used in feature localization routines to obtain the spatial coordinates of an object with high precision and accuracy . this knowledge can also be used to segregate two spots that are closely spaced . @xcite developed a method for automatic detection of diffraction - limited spots separated by less than the rayleigh limit in three - dimensional image stacks under the key assumption that each spot detected in an image is comprised of a finite number of superimposed psf s . using simulated data , thomann _ _ @xcite found that their algorithm can resolve points at sub - rayleigh separation . the localization accuracy approached the nanometer range for signal - to - noise greater than @xmath9 db , and was sub-20 nm for lower signal - to - noise ratios . high accuracy was only maintained for points separated by at least the rayleigh limit , and depended on the relative brightness of the two points . spots of equal brightness could be distinguished at a distance of half the rayleigh limit . for cases where the spots differ in brightness , the resolution limit decreased by up to 50 % . several works address high fidelity tracking in systems where particle trajectories may overlap @xcite , a common problem encountered in using two - dimensional detection to acquire images of three - dimensional objects . imaging using three dimensional detection of diffraction - limited spots separated by greater than the rayleigh limit lifts the degeneracy of particle overlaps and ( dis)appearances . just as the dynamical resolution of point - like objects is limited by the resolution in the feature localization , so also the resolution for volumetric changes of cells or subnuclear regions is limited by the faithfulness of any surface reconstruction being performed . many works address the problem of reconstructing the topography of a surface from sample points in three dimensional spaces @xcite . watertight surface reconstructions , which return a well - defined interior and exterior , are needed for many applications , including computer graphics , computer aided design , medical imaging and solid modeling , but are not guaranteed by many reconstruction methods . basic surface reconstruction techniques can be classified into explicit representations , including parametric surfaces and triangulated surfaces , in which all or most of the points are directly interpolated based on structures from computational geometry , such as delaunay triangulations ; and implicit representations , which solve the problem implicitly in 3d space by fitting the point cloud to a basis function , and then extracting the reconstructed surface as an iso - surface of the implicit function . methods in the second category differ mainly in the different implicit functions used . implicit representations are a natural representation for volumetric data : the marching - cube ( mc)-like algorithm @xcite is currently the most popular algorithm for isosurface extraction from medical scans ( ct , mri ) or other volumetric data @xcite . mc has been applied to subcellular @xcite and cellular @xcite volumetric data using commercially - available software applied to three - dimensional image data obtained by confocal microscopy . mc produces the desired watertight , closed surface and is easy to parallelize , but requires uniform ( over-)sampling , produces degenerate triangles ( requiring post - processing re - meshing ) , and does not preserve features . while the output mesh that mc generates is adequate for visualization purposes , it is far from being suitable for surface or volume quantification in situations of local undersampling due to high curvature . in practice , the data may sample only part of a surface densely , and may suffer from inadequate sampling in other , local regions , either due to non - perfect homogeneous distribution of a fluorescent reporter , or to local high curvature of those regions relative to the imaging ( pixel ) sampling rate . for example , in budding yeast , due to the small cell size and the even smaller incipient daughter cell ( bud ) size , there are local regions in which the curvature may be high relative to the sampling , even when sampled at the highest microscope magnification . existing surface reconstruction algorithms face difficulty if the data contains local undersampling , and will yield holes or artifacts under these conditions . we envision many scenarios where local undersampling is present . the tight cocone method generates a surface representation given a set of sample points , suppressing undesirable triangles and repairing and filling the holes in the surface @xcite . it provides a water - tight reconstruction even when the sampling bounds required by other algorithms are not met . for surfaces that are complex , i.e. that are not single compartments , the tight cocone algorithm will reconstruct the full surface . by decomposing the space and re - running the algorithm , we are able to build more complete volumetric representation of the dividing yeast cells , producing watertight reconstructions of each compartment . in this work , feature finding and tracking in living cells was developed together with surface reconstruction and volume determination . the algorithms are semi - automated to maximize the number of cells that can be analyzed with a minimal amount of subjective manual input . these new methods were applied to investigate budding yeast mitotic spindle dynamics at a high spatial resolution over the entire cell cycle , and the compartment volumes of dividing cells . to analyze the spatial location of mitotic spindle poles in a reference frame intrinsic to the dividing _ s. cerevisiae _ cell , we determined the 3d position of the poles and , in parallel , the cell bounding surface , the plane of cell division ( neck plane ) , and neck plane center , as landmarks . to fluorescently tag spindle poles , we used cells in which the chromosomal copy of spc42 was fused to the red fluorescent tandem - dimer tomato . we additionally used cells in which the chromosomal copy of the g - protein gpa1 was fused to gfp , to visualize the cell surface in the same cells . we imaged asynchronous live - cell populations in three dimensions , yielding images containing up to @xmath10 cells each . and @xmath11 lie in the plane of separating the mother and bud cavity during cell division , called the _ bud neck _ in budding yeast . the @xmath12 coordinate in this definition is positive in the mother cavity and negative in the bud.,scaledwidth=50.0% ] experiments were performed at a temperature of 25 @xmath13c . at this temperature the budding yeast cell cycle is @xmath14 hours . photobleaching restricts the window of time that a single cell can be observed at several nanometer resolution . thus , to capture spindle dynamics over the entire cell division , populations of unsynchronized cells were observed , with each cell in a field - of - view acting as a sample of a specific temporal window of the cell cycle . the spindle poles were labelled with one fluorescent protein and cell periphery was labelled with a different fluorescent protein . the spindle pole trajectories and mother and bud cavity surfaces were reconstructed from fluorescence data as described below . the final result of pole and surface finding is an ensemble of spindle pole trajectories , with the capability for each trajectory to be described in a coordinate system relevant to cell division for that cell . the geometry resulting from such an analysis for a single cell is plotted in fig . [ fig : img_sim : cell_coord ] . in the next section , the algorithms used for extracting spindle pole dynamics and cell surface topography are described . all implementations were written in matlab ( mathworks , natick , ma ) . as described above , knowledge of the expected intensity distribution for a diffraction - limited spot can be used to automatically localize fluorescently - labelled objects that are smaller than the diffraction - limit . the general principle of the localization routine is depicted in fig . [ fig : ff:3d_ff_principle ] . by fitting the full intensity distribution of the object to a three - dimensional gaussian function , the centroid of the spot can be determined to sub - pixel resolution in all three dimensions . to construct spindle pole trajectories from a time series of confocal fluorescence image stacks , an automated three - dimensional feature - finding and tracking algorithm was used . the algorithm locates the centroids of the spindle pole intensity distributions , termed features , to sub - pixel resolution in 3d in a series of successive confocal image stacks acquired at controlled time intervals . it was originally developed as center - of - mass fitting for applications in colloidal systems at 3d sub pixel resolution @xcite , and adapted for application in living cells using gaussian fitting by two of the authors . the tracking algorithm links localized features in sequential image stacks and outputs a time series of coordinates for each feature . the resolution of the three - dimensional localization depends on the signal - to - noise ratio of the images , as will be discussed in detail below . for this study , the localization accuracy is between @xmath15 and @xmath9 nm in 3d . the automated feature finding consists of three main steps : filtering , initial estimation of feature positions , and refinement of position estimates . to suppress image artifacts and noise , which exhibit high spatial frequencies ( small length scale ) , the images are initially filtered using a three - dimensional band - pass gaussian kernel . by choosing a spatial cut - off for the filter kernel close to the characteristic size of a diffraction - limited spot , high frequency noise is attenuated to a level far below that of true features . once the image has been filtered , pixels that are local intensity maxima can be used as initial estimates for feature positions . to further distinguish real features from noise , the integrated intensity is calculated for a sub - volume surrounding each of the local maxima . features for which the integrated intensity is below a cut - off value are excluded from further processing . at this point , the feature position estimates have a spatial resolution equal to the pixel size , typically 180 nm in @xmath16 and @xmath17 and 300 nm in @xmath18 . to refine the feature position , a three - dimensional gaussian function is fit to the intensity distribution within the sub - volume surrounding the initial estimate , using non - linear least squares . the final sub - pixel estimation of the feature location is obtained by a second iteration of gaussian fitting . the second iteration involves first shifting the pixel - spaced sub - volume so that it is distributed around the centroid of the initial gaussian fit @xcite , followed by a second least squares gaussian function fit , using the outputs of the initial fit as parameter estimates for the second fit . the band - pass length scale , the dimensions of the sub - volume used for fitting , the integrated intensity cut - off value , and a characteristic size of the object are used to define a region around the candidate feature in which a second feature can not be located , and are defined by the user . the last parameter is adjusted to allow or reject any overlap between two or more features of interest . these four parameters are set to optimize the fidelity of feature localization across a large number of cells contained in a single field - of - view . since photobleaching degrades the signal of a given feature as it is observed over time , the integrated intensity cut - off can be automatically reduced by a user - defined constant factor for each timepoint in the series of confocal stacks . the tracking algorithm links localized features in sequential image stacks and outputs a time series of coordinates for each feature . the algorithm used here is based on crocker _ et al . _ @xcite and implemented in matlab @xcite . feature coordinates are linked in time by minimizing the total displacement of each of the identified features between successive frames . the search for corresponding features in consecutive frames is constrained to a maximum likely displacement over the timescale between frames . accurate and robust tracking is obtained by tailoring the timescale of image acquisition and the cut - off distance to the particular dynamics of the features being tracked . high fidelity tracking is obtained by applying successive iterations of the tracking algorithm to each data set @xcite . this method ensures that the majority of trajectories over a cell population are captured , and results in data sets biased for trajectories that persist to long times . the cell periphery is visualized using fluorescent reporters that localize to the cell cortex . using the procedure we developed , if the reporter is distributed uniformly over the entire surface , the fluorescence intensity distribution in a confocal image stack can be used to reconstruct the cell surface . automated surface reconstruction is carried out on stacks of images obtained using reporters that label the periphery of living cells , provided the resolution in @xmath19 is sufficient and the label ( dye or fluorescent reporter protein ) distributes uniformly around the cell periphery . key features of the surface shape that are important to cell division , such as the bud neck in budding yeast , may then be extracted . the plane of cell division in budding yeast is called the bud neck : in any dividing cell it is characterized by local deformation of the bounding surface to create two cavities out of one . in addition to the final topology change , local geometry is also modified smoothly during the process of cell division . using the mother and bud cavity fits for each cell , the neck plane position and orientation can be calculated using the methods described here . the automated surface reconstruction procedure we developed consists of five main steps : filtering and thinning , initial estimation of closed bounding surface , refinement of surface estimates by decomposition of the space , calculation of volume of each convex cavity ( here , mother and bud cavity of mitotic cells ) , and , in mitotic cells , determination of position , orientation , and size of the neck separating the mother and bud cavities , which provides a basis coordinate system for cell division unique to each cell . the first step of surface feature finding begins with spatial filtering of the image stack , with the characteristic size parameter of the 3d band pass algorithm corresponding to the surface thickness . using a first estimate of the thickness , one can iterate through different values until the surface and a reasonable cavity size are reproduced with optimum fidelity . the thresholding step must use a threshold value that maximizes the amount of surface included , while minimizing artefacts such as protrusions , small clusters away from the surface , and non - faithful connection between opposite sides of a cavity . typically , the true surface data is comprised of thousands of connected voxels , whereas small noise clusters may span only a few hundred voxels or less , providing an effective means for the identification and removal of small noise clusters . the thresholded image stack is then sliced in @xmath20-plane images , each of which is `` thinned '' to a minimal collection of points that preserves holes and general shape , to allow for a full surface reconstruction . thinning was used instead of skeletonization which is less suited for minimizing the surface data as it sporadically produces branches . the thinning algorithm transforms an entity within the image into a single pixel thickness entity @xcite . the algorithm yields good results on slices near the equator of the surface to be analyzed , where it reduces a thick shell to a 1 pixel - wide shell . an example on one focal plane of data acquired from the surface reporter imaging in budding yeast , before and after the thinning step , are shown in fig . [ fig : meth_surf3 ] ( a - b ) . on nearly - flat curved surfaces , which are manifest in slices through the top or bottom of a cell , the thinning algorithm has the undesired effect of reducing a uniform disk to a single or branching lines . by additionally slicing the image stack into sets of @xmath21 and @xmath22 planes , and processing these sets of planes with the thinning algorithm , we eliminate these artifacts specific to the slicing direction . in these new slices , `` top '' and `` bottom '' caps are instead part of equatorial profiles . the result produces good performance , defining the top and bottom of the cell as a collection of points in a contour . the estimated minimal set of points describing the surface is then constructed by merging the results from the @xmath20 , @xmath21 and @xmath22 slice directions . we assemble the final cloud of points by requiring that for a point in space to be retained , it must appear in at least two of the slicing directions . the resulting 3d thinned version of the thresholded image stack is composed of a cloud of points centered approximately in the middle of the original thick surface of the image stack . again small disconnected clusters of points can be removed from the thinned surface by requiring a certain minimum size of connected points . the output of this discrete cloud of points estimation step is used as input to the voronoi - based algorithm to obtain useful surface representation , which can then be used to quantify volume . to define the volume , the scattered point samples of the surface need to be turned into a proper watertight surface mesh . in cases where the signal is strong and uniform , an accurate , arbitrarily - shaped tessellated surface can be obtained via a full surface reconstruction . we achieve this using the tight cocone software @xcite applied to the minimized discrete surface points . the tight cocone algorithm takes a point distribution in 3d and attempts to reconstruct a water - tight surface described by those points . the algorithm is robust to local undersampling as intended , but as it is sensitive to noise , we supply the filtered data . the output is a list of triangles that comprise the reconstructed surface . as a proof - of - concept example , the result on one budding yeast cell for which the surface is thus tessellated is shown in fig . [ fig : meth_surf3 ] . the determined surface is remarkably faithful to the actual cell s topography , reproducing bumps , curves , and the typical smooth curvature transitions at the bud neck . from a reconstructed surface , quantities such as volume may then be computed . if the shape is a single convex hull , the volume can be computed simply , by using a well - known algorithm . if the complete surface is not a single convex hull , this approach will not robustly provide the volume . for example , in budding yeast , the bud cavity can protrude significantly out of the mother , with a concave neck region linking the two cavities . however , individually , the mother and bud cavities are convex . we effectively decompose the complex multiple - hull surface into individual convex hulls , for each of which compartment the volume can be determined . in principle , the negative curvature of the neck region could be used to obtain the orientation and location of the plane separating the mother and bud cavities . in practice , even though the periphery is well - labeled so that the whole surface can be reconstructed , undersampling prevents determination of the local curvature everywhere . moreover , in small budded cells the neck curvature may only just change sign at the transition region from mother to bud . instead , we exploit other information to uniquely identify the tessellated surface data in this region between the two cavities . in our dividing cells , we subdivide the surface into sections by performing a fit of two ellipsoids to the minimized surface points . for each vertex of the closed bounding surface , we then compute the shortest distance to the fitted ellipsoids . the computed proximity is used to segregate the vertices into two subsets : those assigned to the mother cavity , and those assigned to the bud cavity . after this assignment step , we perform the water - tight surface reconstruction on each segregated point set . this provides individual convex surfaces which are used for robust calculation of volumes for each of the two cavities . following this step , those triangles in the original full surface reconstruction having at least one vertex belonging to each of the mother and bud sets are assigned to the neck region , and are used to reconstruct the surface of the neck region . all of these steps are carried out for each dividing cell . the volume of a convex hull can be computed simply , as follows . considering any point on the surface of or inside the volume , the volume is calculated as the sum of all pyramids whose base is one of the triangles forming the surface and whose summit is that point . these pyramids constitute a complete set , containing the volume enclosed by the surface , and the sum of their volumes is the total volume of the convex hull . from the minimized points forming the neck surface , we extract geometric characteristics of the neck by fitting a circle ( or ellipse ) to these points . this yields the neck plane location ( including the center ) ; circumference ; and , via its normal , orientation with respect to the imaging frame of reference . knowledge of this plane provides a reference frame for cell division unique to each cell . as can be seen in fig . [ fig : meth_surf3 ] ( d ) , this automated procedure results in faithful description of the orientation and position of the neck plane between the mother and bud cavities . the results obtained show the suitability of the method for a correct representation of the target object . the trajectories of other point - like features obtained simultaneously with the surface in each cell can now be transformed into their cell - unique coordinate system , and therefore examined relative to the neck origin . together , the spindle position and surface mapping could additionally be used in tandem to define a wealth of physical observables . measuring the dynamics of the mitotic spindle in a coordinate space defined by the plane of cell division , over a population of cells , would reveal dynamical and shape cues related to stages of the cell cycle not quantifiable precisely by other methods . specifically , spindle orientation , and translocation with respect to the neck plane , could be obtained automatically . the orientation of the normal vector to the neck plane , @xmath23 is determined from the fitting as described above , and its sign can be assigned as desired : we here define @xmath24 as extending into the mother cavity . a spindle vector @xmath25 may be defined as the vector spanning from the pole proximal to the bud to the pole distal to the bud . spindle orientation is then parametrized by the angle @xmath26 between @xmath25 and @xmath23 , given by @xmath27 , where @xmath28 . for each cell , the spindle orientation and its angular excursions may be calculated over the observation window . the distance of the spindle poles from the neck plane may be computed by constructing the vectors @xmath29 , representing the displacement of each spb located at point @xmath30 from the neck centre at point @xmath31 . the distance , given by @xmath32 , is positive for all @xmath30 in the mother cavity , and negative for all @xmath30 in the bud cavity , using our definition of the @xmath23 direction . accurate segregation of aberrant spindle positioning behavior from that of normal cells would require knowledge of the spindle size and position relative to the mother and bud cavities . such studies , achievable at high resolution in cells in sc medium with both poles and surfaces labelled , will be enabled by these tools . the ability to calculate compartment volume might have other uses . not all volumes one wants to quantify are conveniently regular . chromosome territories , nucleoli , and golgi bodies are all examples of compartments with complex distributions or irregular outlines , occupying in higher eukaryotes approximately the same volume as the yeast cells under study here , a few tens of femtoliters @xcite . methods to study such compartments in a statistically meaningful manner have been limited and are generally complicated by the fact that estimating the volume of such exclusion zones requires calculating the volume of irregular domains . the accuracy of feature finding and tracking in cells is influenced by many factors . signal - to - noise must be maximized with well defined intensity distributions for each feature , spanning multiple voxels so that the nyquist criterion is satisfied . the hardware , signal - to - noise , and sampling requirements limit the achievable temporal resolution . investigation of fast dynamics in the cell requires some trade - off between temporal and spatial resolution . a detailed assessment of the factors influencing the performance of localization algorithms to determine the limits of the spatial resolution greatly aids in the design and interpretation of feature finding experiments . a simulation of diffraction - limited spots that is relevant to the experimental conditions requires knowledge of the true psf of the optical system used for the experiment . the theoretical description of a confocal psf is only relevant in the ideal aberration - free scenario , and it is necessary to measure the real psf of the microscope . measurement of the psf will reveal any aberrations that may be present in the optical system . the image of an object that is smaller than the diffraction - limit closely approximates the true psf of the microscope . in an optical system , a diffraction - limited object acts as a low - pass spatial filter with a high cut - off frequency relative to that of the otf . the cut - off frequency increases as the inverse of the object size . the true psf is the image of an infinitesimal object , i.e. one with an infinite bandwidth . however , in practical imaging conditions , background and noise signals often mask any contribution of high frequency components to the psf , so that the effect of a small but non - zero object size on the psf is negligible . therefore subresolution fluorescent beads may be used to measure the psf . measurement of the psf for our optical system is described in the supplementary materials @xcite . to simulate all aspects of image acquisition realistically , the background or bias signal and the readout noise from the ccd camera were also measured , as also described in @xcite . in order to determine the accuracy and precision associated with the feature localization algorithms used in this hybrid method , the localization algorithms were applied to computer generated confocal images of spindle pole bodies . our procedure for generating these simulated images , using parameters drawn directly from our hardware , signal - to - noise , and 3d sampling characteristics , is described in @xcite . for each simulation run , an image was generated at a given signal level and with a given displacement between the two diffraction - limited spots in a prescribed direction in three - dimensional space . noise characteristic of the yeast medium and camera ( as measured ; @xcite ) was added to the image , and the feature localization algorithm was applied . at each signal level and spot separation tested , 500 iterations of feature finding were performed with a new noise profile generated for each iteration . all simulations were performed in matlab . feature localization parameters were set to the values used for live cell trials . the feature coordinates , as determined by the localization algorithm , were then compared to the real coordinates to obtain an estimate of the localization error over a range of signal - to - noise values and spot separations . knowledge of the true spot locations allows for an estimate of both precision and accuracy of a given measurement . the deviation between the actual position and the localized position was derived in this way . the result for a point - to - point displacement typical of a metaphase spindle , over a range of signal - to - noise levels , is shown in fig . [ fig : img_sim : error_vs_rsn ] . anaphase spindles have greater separation and even less vector component along the @xmath18 direction , so suffer less from localization error than than do metaphase spindles . budding yeast cells are small , rendering subcellular quantitation challenging . nevertheless , the spindle poles in these cells are separated by @xmath33 in the early phase of cell division ( metaphase ) , and by larger distance later in the cell cycle . thus the metaphase spindle is the more stringent test of true subpixel and aberration - free resolving of the spindle pole features from one another . in our applications studied here , metaphase and anaphase , the spindle pole psfs in our optical system do not overlap . using the simulations we had developed , we tested the effect of overlap at distances of the same order as , and smaller than , our metaphase spindle length , to validate that effects of overlapping of features need not be taken into account at these distances , and determine at what separations such effects play a role . interestingly , we found that when filtering is used , the filtering itself is a source of small overestimation of the separations at distances @xmath34 . for such separations , spatial band - pass filtering attenuates the summed intensity distribution in the region between the two points , and forces the least - squares fitting routine to localize the centroid of each intensity distribution away from the overlap region , resulting in overestimation of the point - to - point separation . as overlap increases , the centroids of each feature are forced back towards the overlap region , initially decreasing the estimated separation , until significant overlap has occurred at @xmath35 , and the two spots are difficult to distinguish ( see figs . [ fig : img_sim:2poinintdist ] , [ fig : img_sim : msep_vs_asep ] ) . for automated feature detection with live cell image data , spatial band - pass filtering is often necessary . real data is subject to many sources of noise , and accurate automated segregation of the real features of interest from noise is often not possible without spatial filtering . therefore , although our feature separations here well exceed this limit , the systematic bias demonstrated here will always contribute to the localization error for closely - spaced features . although small , this effect must be considered in the interpretation of the results of a feature finding experiment under such conditions . propagation of localization error in the separation of two points was studied in order to determine its role , if any , in the observable @xmath36 , the spindle pole - pole separation @xcite . comparison of measured error with propagated error for point - to - point separation showed that for large separations , the covariant terms are negligible ( see fig . [ fig : img_sim : errorpropcompare ] ) . as the two features approach one another to within less than @xmath37 , the covariant terms become significant and they must be included for an accurate estimate of the error . we determined that these two effects must be taken into account when the distance between two spots is such that their point spread functions overlap @xcite . in our observable , budding yeast spindle length in metaphase and anaphase , the spindle pole - pole distances we are measuring always exceed these limits . we used budding yeast as a model system in which to study mitotic spindle dynamics . live unsynchronized cell populations expressing tdtomato - tagged spc42 , a protein in the spindle pole body ( spb ) protein complex , and the surface reporter gpa1-egfp , were imaged by confocal fluorescence microscopy . spb s labelled with a red fluorescent protein variant ( tdtomato29 @xcite ) at the protein spc42 are point - like features which were localized and tracked using the methodology described above . surface data was analyzed using the surface fitting described above . .,scaledwidth=48.0% ] all yeast strains used were derived from strain by4741 @xcite . media used for yeast culture are described in @xcite . yeast genetic manipulations were carried out as described in @xcite . for microscopy , yeast strains were incubated in sc at 25@xmath13c until early logarithmic phase ( @xmath38 normalized optical density at @xmath39 nm ) . to obtain the desired cell number density for imaging , 1 ml of suspended cells were pelleted for 15 s at @xmath40 , washed twice with @xmath41 of lactate medium , then re - suspended by gentle mixing 100 times with a pipet , and @xmath42 of culture was pipetted onto a sc - infused agar pad on a slide and covered with a glass coverslip . cells were imaged at 30@xmath13c in stacks of 21 focal planes spaced @xmath43 apart , spanning the entirety of the cells in the axial direction , at 10 s per stack , using a custom multi - beam scanner ( @xmath44 pinholes ) confocal microscope system ( visitech ) mounted on a leica dmirb equipped with a nanodrive piezo stage ( asi ) and @xmath45 and @xmath46 solid state lasers controlled by an aotf . images were acquired with a em - ccd camera ( hamamatsu ) using a 63x 1.4 na objective . from the images , we computed the positions of the spindle pole bodies at three - dimensional sub - pixel resolution @xcite by fitting the intensity distribution to a three - dimensional gaussian function ( fig . [ fig : ff:3d_ff_principle ] ( b ) ) , obtaining a spatial resolution of @xmath47 in 3d for each pole . all analysis was carried out using custom scripts in matlab ( mathworks ) . for imaging surfaces , yeast strain of exactly the same genotype with a g - protein labeled as a gfp clone was obtained ( life technologies , @xcite ) . the g - protein , gpa1-egfp , localizes to the cell cortex and labels the whole cell periphery @xcite . for all analysis , data was collected from cell populations representing four independently - derived yeast strains . for combining of position data from the two channels , chromatic shift between the @xmath48 and @xmath49 excitation channels was measured and corrected for prior to combining results ( see @xcite ) . we discovered that for yeast cells imaged in lactate medium , exposure to laser irradiation for image stack collection at 5 second intervals results in a phototoxic effect that is not alleviated by increasing the quantum efficiency of the fluorescent reporters . the effect causes cells to arrest at the transition from metaphase to anaphase . this effect was confirmed by observing cells in different media types , and varying exposure times for prolonged periods under sealed coverslips . literature - reported values for the rapid phase of anaphase b in budding yeast demonstrate that this rapid phase is a slowly - varying process compared to the image stack acquisition rate we were using . therefore , our acquisition scheme should capture the transition from pre - anaphase to anaphase . specifically , in observation of an asynchronous budding yeast cell population over a finite 20-minute observation window , representing @xmath501/5th of the cell cycle , the transition should be observed for @xmath5030 cells in an analyzed population of 160 total cells . initially imaging in lactate medium , we observed a major discrepancy from the expected biological phenotype : not a single cell traversed to anaphase while under observation . to investigate this and determine carefully in what regime the cells can be observed without any photon - absorption - induced perturbation , wild - type yeast strains expressing the spindle pole reporter spc42-tdtomato were prepared for microscopy as discussed above using lactate medium . a small population of asynchronous cells were imaged under a sealed coverslip for 30 minutes at 20 second exposure intervals . six different fields of view were examined with approximately ten cells per field of view . no qualitative evidence of anaphase spindle elongation was observable in any of the cells . another population of cells was cultured , washed and imaged in a minimal medium , synthetic complete ( sc ) , which consists only of the 22 amino acids for yeast together with sources of carbon ( dextrose ) and nitrogen ( peptone ) @xcite . four different fields of view were examined with approximately ten cells per field of view . the cells were imaged for 30 minutes at 5 , 10 , and 20 second exposure intervals . in this minimal medium , a proportionate ( expected ) number of cells displayed anaphase spindle elongation , for all the exposure intervals studied . furthermore , the labelled spindle poles were observed to maintain signal intensity longer when imaged in sc medium than in the lactate medium , further indicative of fewer oxygen free radicals being created from photon absorption . following these initial studies , the spindle dynamics of cells incubated and imaged in sc medium was investigated in detail , using the protocol for quantitative image acquisition we derived from our studies . to minimize phototoxic effects , the acquisition interval was selected to be 10 seconds , and the total acquisition length to be 20 minutes . a population of unsynchronized wild - type cells expressing the spc42-tdtomato reporter was incubated , washed , and imaged in sc medium . a confocal stack was collected with the 491 nm excitation line for each time point ( @xmath51 sampling rate ) , with the exposure time set to @xmath52 . within a given field - of - view , only cells for which two spindle poles could be detected within our resolution limits were analyzed . signal - to - noise levels were typically between 7 and 11 db . the spindle pole dynamics were analyzed in all cells in three - dimensions . in these data , those cells with two poles were segregated from the data . for each time - point in these cells , the displacement between the two poles was computed , to construct the time evolution of spindle length . for cells in metaphase and early to mid - anaphase , the separation between spb s corresponds to the spindle length . an additional level of segregation based on spindle length can thus be performed to yield separate metaphase and anaphase subpopulations . for cells in the latest stage of anaphase , the elongated spindles begin to buckle and then disassemble as a mechanical entity , so that in this latest phase of anaphase , the spb separation does not reflect the true , curvilinear , intact spindle length . the measured spindle pole separation versus time obtained following analysis of the resulting images is plotted in fig . [ fig : results : sc_wttraj ] . for representative cells from the total imaged population , individual spindle - length trajectories across the entire imaging time are displayed . as a function of time , in a cell traversing anaphase . ( b ) results are displayed for 3 metaphase cells showing representative bipolar spindle behaviors from a total population of @xmath53 analyzed cells each of wild type ( blue triangles ) , kip1@xmath54 ( orange stars ) , and cin8@xmath54 ( green circles ) . ( c - e ) representative anaphase behaviors of ( c ) , wild type cells ; ( d ) , kip1@xmath54 cells ; and ( e ) , cin8@xmath54 cells.,scaledwidth=48.0% ] these data obtained from cells imaged in sc medium demonstrate the full range of stereotypical dynamical phases that are characteristic of normal progression through the cell cycle . there are represented the approximately steady - state pole - pole separation at @xmath33 characteristic of yeast metaphase ( sometimes called pre - anaphase)(fig . [ fig : results : sc_wttraj ] ( b ) ) . there are the directed , apparently irreversible rapid spindle elongation , characteristic of yeast early anaphase @xcite , in which pole - pole separation increases rapidly from @xmath55 to @xmath56 over a duration of several minutes , followed by slower , less robust spindle elongation ( fig . [ fig : results : sc_wttraj ] ( c ) ) . there are additionally cells with initially long spindles , in late anaphase , that exhibit uncorrelated pole - pole motion characteristic of total spindle disassembly once anaphase . using these tools and in sc medium , assignment of these cell cycle phases in an unsynchronized population may be done via automated segregation of individual traces based on their dynamical behavior . since in budding yeast , chromosome condensation and a metaphase plate are not visible by microscopy , this segregation provides a valuable intrinsic metric for cell cycle state . in their performing of useful work in the cell , it is expected that the ensembles of kinesin-5 motors exhibit collective dynamics under nonequilibrium conditions . this should manifest itself in the observable @xmath36 . for cells in metaphase , the spb separation exhibited fluctuations superimposed upon a nearly constant , slowly increasing , function time . these fluctuations were well above the resolution limits of our feature finding of @xmath57 nm , as the signal to noise level over the imaging duration remained @xmath58 . for cells in anaphase , the motor activity at the spindle mid - zone is increased @xcite . due to their intrinsic directionality on the antiparallel mts at the spindle mid zone , the extensile , coherent dipole character of the motors becomes manifest as directed , biased motion . at the scale of the whole spindle , this results in elongation . the collective dynamics of the motors gives rise coherent motion over longer timescales than the individual motors typically exhibit @xcite . with the precision we obtain for the pole - pole distance in 3d , we can determine whether fluctuations also appear superimposed on the directed , biased anaphase motion , as they do on the metaphase motion . upon observation of the motions with the detail enabled here , the anaphase motions were seen to be coherent , absent the tens of nanometer scale fluctuations we observed in metaphase . to investigate the specific contributions of mitotic motors to spindle dynamics , genetic perturbations of the mitotic motors were induced in budding yeast . imaging and analysis of spindle dynamics using these tools was carried out for cells in which the gene for one or the other of the two kinesin-5 mitotic motors , @xmath59 and @xmath60 , was deleted . aside from these single - gene deletions , the genotypes were identical to that of the unperturbed cells studied above . populations of cells in which one kinesin-5 motor was deleted were prepared with the @xmath61 fluorescent label . these @xmath62 or @xmath63 cell populations were cultured and imaged in the sc medium as described above . data for the measured spindle pole separation versus time , obtained following analysis of the resulting images , for representative cells from the total imaged populations , are plotted in fig . [ fig : results : sc_wttraj ] . the metaphase spindle length fluctuations displayed in fig . [ fig : results : sc_wttraj ] ( b ) do not differ appreciably between wild - type and mitotic motor mutant populations . many of the cells are observed to pass through anaphase spindle elongation in both the kip1 and cin8 deletion strains . the nature of the anaphase motion in kip1-deleted cells , i.e. those carrying only the cin8 kinesin-5 , is very similar to that in wild type cells , aside from slower spindle pole separation in late anaphase . by contrast , the cin8-deleted cells , i.e. those carrying only the kip1 kinesin-5 , move more slowly through anaphase , consistent with what has been observed previously @xcite . additionally , this population also shows much greater variability in overall rate of spb - spb separation from cell to cell ( data not shown ) . the nanometer - scale - resolution enabled by the tools we developed here permit more detailed examination . upon further scrutiny , we see that the data for anaphase in @xmath64 cells show some stop - start stuttering , where the extensile motion is apparently persistent for several tens of seconds , punctuated by periods of motion in the opposite direction . this oppositely - directed , contractile - like motion is not apparent in the anaphase motions of wild - type and @xmath62 cells . it could arise from either brief , passive partial collapse of the spindle , or coherent ( persistent ) motion in the opposite direction ; that is , active contractile motion after directional switching . directional switching has recently been observed for the kip1 motor _ in vitro _ under certain conditions @xcite . we have recently used these context - rich tracking tools to examine the statistical mechanics of the fluctuations during metaphase for a large population of cells , in the context of models for the non - equilibrium activity of motors / microtubule ( de)polymerization @xcite . it will be interesting to examine the statistical mechanics of the coherent motions during anaphase , for populations of non - genetically perturbed cells and populations of cells with mitotic kinesin-5 molecular motors deleted , enabled by these methods , and to compare these observation to _ in vitro _ experiments once _ in vitro _ systems are developed in which controllable _ ensembles _ of kinesin-5 motors can be studied @xcite . mother and bud compartment volumes of dividing cells were calculated using the method described above applied to the thinned data . in fig . [ fig : results : mutant_volumes - dists ] , the bud cavity volume is plotted versus the mother cavity volume . each data point corresponds to a separate wild - type cell . as expected , the bud cavity was always smaller than the mother cavity @xcite . in fig . [ fig : results : mutant_volumes - dists ] , the mother and bud volume distributions over the cell population are plotted . each of the mother and bud volume distributions were fit to a gaussian function . there is a degree of variability in the volume of the mother cell and in the volume of the daughter cell within a cell population , as was observed in early quantification using area as a proxy for volume @xcite . despite the variability , the distributions are bounded from above and below , due to known mechanisms of the cell cycle . the distribution of mother cells is bounded on the lower side by a minimum cell size threshold for replicative division ; and on the upper side by the fact that the time from one division to the next is compressed in cells that have already divided , and by the fact that during mitosis , growth of the mother cell is minimal as most of the increase in mass goes into bud growth @xcite . ( red ) or @xmath65 ( blue ) . the line at @xmath66 is drawn for reference . bottom : probability distributions of bud and mother cavity volumes for populations of pre - anaphase wild - type cells ( dark green , bud ; light green , mother ) and for populations of cells with mitotic kinesin deletions ( @xmath67 : bud , light red ; mother , dark red ; @xmath65 : bud , dark blue ; mother , light blue).,scaledwidth=48.0% ] cell volumes quantitated for the motor deletion strains are displayed in fig . [ fig : results : mutant_volumes - dists ] , superimposed on the results for wild - type cells . the data show that @xmath67 cells have mother and bud volume distributions practically unchanged from those of wild - type cells . by contrast , the volumes in the @xmath65 population are markedly larger , for both the mother and bud cavities , than in wild - type and @xmath67 populations . moreover , the distribution of @xmath63 volumes across the population is much broader for both mother and bud , and the mother and bud distributions are no longer distinguishable . the mean and standard deviations of mother and bud cavity volumes obtained from gaussian fits to all the distributions are collected in table [ tab : volumes ] . .mean @xmath68 and standard deviation @xmath69 of the distribution of mother and bud volumes over populations of @xmath1 cells of wild - type cells or with one or the other of the mitotic kinesin-5 motor proteins deleted . [ cols="^,^,^,^,^,^",options="header " , ] the differences observed between wild - type and motor - deleted populations on the large mother and bud cavity volumes suggests a potential relationship between the control mechanism for dynamics of chromosome segregation , in which the kinesin-5 motors are involved , and that for cell size . most striking is the effect of @xmath60 deletions on cell volume . it is possible that the large volumes observed in the @xmath65 population were a result of a delay in the cell cycle for this population during a stage of cell growth : if chromosome segregation in the mitotic spindle is delayed , the cell may continue to produce and partition membrane and mass to the bud , resulting in a bud much larger than normal ; eventually , the distribution stabilizes in homeostatic equilibrium after multiple generations . nevertheless , there is a remarkable correlation , synchrony . given the synchrony already observed between sister chromatid separation and onset of rapid spindle pole separation , specifically at the molecular level by increased coherent kinesin-5 motion in the spindle midzone . this synchrony is now known to be due sharing of the _ same _ signaling elements across the two processes @xcite . it would not be surprising if cell wall growth were also shown to be in synchrony through some shared feedback mechanism with mitotic motors on the spindle . since cell growth is coordinated with the cell cycle @xcite , the bud - to - mother volume ratio serves as a complementary metric for cell cycle progression that at a larger spatial scale than the spindle . to explore this further , we demonstrated here use the tools we developed in tandem . in fig . [ fig : results : volratio_vs_meanl_overlay ] , the bud - to - mother volume ratio is plotted against the mean sbp separation @xmath70 . for wild - type pre - anaphase cells , the bud - to - mother volume ratio is approximately proportional to @xmath70 , indicating a correlation between cell growth and mitotic progression . the data suggest that spindle length and bud - to - mother volume ratio are coupled : there is a constant ratio between nuclear size and cell size . there may be global mechanical signals that tell a cell or organelle about its overall shape . the coupling between mitotic progression and cell growth appears to be a fundamental property of wild - type pre - anaphase cell populations @xcite . although a link between nuclear size and cytoplasmic volume has been suspected for many years , recent studies in yeast have introduced a tractable genetic system in which this question could be answered . [ fig : results : volratio_vs_meanl_overlay ] also plots the bud - to - mother volume ratio versus @xmath70 for the motor - deleted populations . in the @xmath67 cell population , the bud - to - mother volume ratio versus @xmath70 follows the same trend as for wild - type cells . for the pre - anaphase @xmath65 population , the bud - to - mother volume ratio is dramatically less coupled to @xmath70 as compared to wild - type and @xmath67 cells . in some @xmath65 cells , the bud has grown larger than the mother , exhibited by the values of the bud - to - mother volume ratio greater than one . removal of the cin8 motor may perturb components of the control system , altering the coupling between chromosome segregation and cell size control . this can help to explain how the bud cavity may become even larger than the mother cavity . it suggests that if there are internal mechanical signals that tell a cell about its overall state at a larger length scale , these signals are perturbed when the composition of certain force - generating elements of the mitotic spindle are perturbed . although cell size control in yeast has been explained for decades in terms of a minimum cell size before entry into mitosis @xcite , many questions remain . for example , it is unclear how yeast which lack lamins and lamin - associated proteins adjust nuclear volume in response to changes in cytoplasmic volume . in any organism , the mechanism by which the upper limit to nuclear growth is established is unknown . ( red ) or @xmath65 ( blue ) deleted.,scaledwidth=53.0% ] the results presented here serve as a proof of principle of the methods developed to track spindle pole bodies with high accuracy in full 3-d space , and faithfully reconstruct budding yeast cell surfaces and volumes . we have presented of a novel application of high - resolution position finding in conjunction with computational geometry techniques to reconstruct surfaces of cells as small as a few micrometers , to locate and track dynamic entities at their address in a living cell , and demonstrated the use of this method in capturing subcellular dynamics . besides imaging features at high resolution in context - rich fashion , our technique also makes possible measurement of the volume and mapping of the surface of small individual cells or subnuclear regions for which the signal from the bounding surface may suffer from local reporter heterogeneity or high underlying surface curvature . the ability to calculate volumes and shapes in situations of low sampling of surface points is likely to become more important as we develop more sophisticated models of how the interior of the cell and the interior of the nucleus are organized , and how they deform and transduce forces . the authors thank professor tamal k. dey , department of computer science and engineering , ohio state university for providing us surface mesh generation software , and dr . jeffrey n. strathern , nih / nci , for providing the yeast strain carrying the spc42-tdtomato fusion protein . 54 natexlab#1#1bibnamefont # 1#1bibfnamefont # 1#1citenamefont # 1#1url # 1`#1`urlprefix[2]#2 [ 2][]#2 , , , * * , ( ) . , , , , , * * , ( ) . , , , , * * , ( ) . , , , , , * * , ( ) . , , , , , , , , , * * , ( ) . , , , , * * , ( ) . , , , , , * * , ( ) . , , , , , , , * * , ( ) . , * * , ( ) . , * * , ( ) . , * * , ( ) . , , , * * , ( ) . , , , , , * * , ( ) . , , , , * * , ( ) . , , , * * , ( ) . , , , * * , ( ) . , , , * * , ( ) . , * * , ( ) . , * * , ( ) . , , , , * * , ( ) . , , , , * * , ( ) . , , , , , , , * * , ( ) . , * * , ( ) . , in _ _ ( ) , pp . . , , , , , * * , ( ) . , , , in _ _ ( ) , pp . . , , , , * * , ( ) . , in _ _ ( , , ) , pp . . , in _ ( ) , pp . . , * * , ( ) . , , , , * * , ( ) . , , , * * , ( ) . , , , , , , , , , , , * * , ( ) . , , , , , , * * , ( ) . , , , , , , , * * , ( ) . , , , _ _ ( , , ) , ed . _ ( , ) . , , , , , , , * * , ( ) . , * * , ( ) . , * * , ( ) . , , , , , * * , ( ) . , , , , , , , , , , * * , ( ) . , , , , , , , * * , ( ) . , , , , , , , , , , * * , ( ) . , , , * * , ( ) . , , , , , * * , ( ) . , * * , ( ) . , , , * * , ( ) . , , , , , , * * , ( ) . , in _ _ , edited by ( , ) , chap . , pp . , ed . _ _ , ( ) . , _ _ ( , ) , ed . * supplemental materials : automated three - dimensional single cell phenotyping of spindle dynamics , cell shape , and volume * to measure the true psf of the microscope , sub - diffraction beads were imaged in 3d with a confocal point scanner ( vt - eye , visitech international ) attached to an inverted microscope ( leica dm4000 ) with a @xmath71 oil - immersion objective ( leica microsystems hcx pl apo , numerical aperture of 1.4 ) at wavelengths @xmath45 and @xmath46 . sub - resolution @xmath72 diameter beads ( tetraspec , molecular probes , eugene , or ) were imaged at @xmath73 exposure time . coverslips were coated with @xmath74 poly - d - lysine to adhere beads to the coverslip . the beads remained immobile throughout the experiment . the concentration of beads was sufficiently low that they were spaced far apart , and only 15 to 20 beads were visible in a given field - of - view ( @xmath75 ) . for five different fields - of - view , a set of confocal image stacks was acquired at a spacing of 0.2 @xmath76between confocal planes . typically , 40 stacks were acquired per set . to estimate the noise - free psf , the intensity distribution of each bead in each of the image stacks was localized to sub - pixel resolution using the feature finding algorithm described below . all intensity distributions for a single bead in a set were then averaged with their centroids aligned . since noise is not temporally or spatially coherent , this averaging dramatically reduces the noise signal . @xmath20 and @xmath21 profiles of the psf obtained in this manner for the @xmath77 excitation channel are shown in fig . [ fig : psf ] . plane , and ( b ) the centre @xmath21 plane , of the three - dimensional point spread function measured by acquiring , localizing , and averaging intensity distributions of 100 nm diameter fluorescently labelled beads . the psf for the @xmath48 channel shown here . the psf for @xmath49 channel is similar ( see text).,scaledwidth=95.0% ] the measured point spread function was symmetric about the @xmath2 , @xmath78 and @xmath19 axes , and showed no significant optical aberrations . the measured psf s were fit to a three - dimensional gaussian distribution @xmath79 $ ] , to obtain a model psf for simulations . reported values are the mean and standard deviation of the gaussian fit parameters , over all five different fields - of - view . the standard deviations of the gaussian psf obtained for the gfp ( @xmath48 ) channel were @xmath80 pixels , @xmath81 pixels , and @xmath82 pixels ( @xmath83 , @xmath84 , and @xmath85 ) in the @xmath2 , @xmath78 , and @xmath19 directions , respectively . for the tdtomato ( @xmath49 ) channel , the measured standard deviations of the gaussian psf were @xmath86 pixels , @xmath87 pixels , and @xmath88 pixels ( @xmath89 , @xmath89 , and @xmath90 ) in the @xmath2 , @xmath78 , and @xmath19 directions , respectively . the readout noise can not properly be determined by calculating a histogram for all pixel intensities across a single image because fixed - pattern noise , whose sources are pixel - to - pixel variations in dark current and non - uniformities in photoresponse @xcite , add additional spread to the intensity distribution . the background signal in the absence of fixed - pattern noise was obtained by using the intensity recorded for a single pixel over many measurements . to accurately reproduce the background present when imaging live cells , a sample of imaging medium ( sc ) was placed in the microscope during the measurements . the objective lens was focused to a point in the media just beyond the coverslip , and 3000 images of a single focal plane were acquired using the shortest possible exposure time ( @xmath91 ms ) . this was performed under three different conditions . in the dark field condition , all shutters were open but no illumination source was used . for gfp and tdtomato illumination conditions , the 491 nm and 532 nm excitation lasers were respectively turned on . since no florescent reporter is present in the sample , the measured intensity distributions represent an estimate of the detector background signal including scattered light effects , as well as ccd readout and dark current noise . because the relevant units for describing noise in the photon detector are electrons , the signal is converted from adu s to electrons in the ccd by multiplying by the ccd gain factor of 5.8 electrons / adu and dividing by the em gain factor used in the acquisition . detector noise from dark current is expected to vary as @xmath92 , where @xmath93 is the dark current and @xmath94 is the exposure time . for the camera used , @xmath95 electrons / pixel / sec @xcite , so @xmath96 electrons , and the noise from dark current is expected to be negligible . readout noise should be gaussian distributed with a variance of @xmath97 electron @xcite . histograms were calculated for a single pixel at different regions over the field - of - view for each illumination condition . for the dark field , 491 nm illumination , and 532 nm illumination , no appreciable change was observed between the histograms from different regions across the field of view . the intensity distribution for each of these conditions follows a gaussian profile , but with a small exponential tail . this exponential tail may be explained by multiplicative noise in the gain - register of the em - ccd , which will produce an exponential distribution of pulse heights with many small pulses and few large ones @xcite . the image of a diffraction - limited spot was constructed by first generating a sphere of @xmath98 diameter in a high resolution object space . the object space containing the sphere is a three - dimensional grid of voxels with dimensions of 29 nm in @xmath2 and @xmath78 , and 50 nm in @xmath19 . the choice of object space pixel size is a compromise between approximation of a continuous space , which requires small pixel sizes , and available computer memory , which limits the size of arrays that can be convolved in a practical time frame . a gaussian approximation to the microscope psf was also generated in the object space using the parameters found earlier . the gaussian function used to generate the psf was not normalized and the amplitude was set to 1 unit . convolution of the object and psf was performed by two - dimensional fast fourier transform of each of the arrays , multiplication in the frequency domain , and finally , inverse fourier transformation ( fig . [ fig : img_sim ] ( c ) ) . to avoid circular convolution , the image and psf arrays were appropriately zero padded prior to fourier transformation . convolution of the object with the psf results in the intensity distribution of a diffraction - limited spot in the continuous image space , the magnitude of which is interpreted in terms of a number of photons . to generate a realistic image that is comparable to experimentally acquired images , it is necessary to sample the continuous image space intensity distribution into a discrete distribution collected by the ccd detector . to do this , the image space intensity distribution is resampled to a pixel grid , with each pixel representing an individual ccd element ( fig . [ fig : img_sim ] ( d ) ) . to simulate the spindle poles , two diffraction - limited spots are produced in the image . the ccd grid has voxel dimensions of 174 nm in @xmath2 and @xmath78 , and 300 nm in @xmath19 , to mimic the experimental conditions . to construct the ccd voxel distribution , the continuous image space distribution was sampled at a number of focal planes in the @xmath19-stack . finite depth - of - field was explicitly accounted for at each @xmath19-slice by weighting the image space matrix by a gaussian function with an amplitude of 1 and a standard deviation of @xmath99 pixels , centered on the particular @xmath19-slice . the intensity distribution in the focal plane was then constructed by integrating the depth - of - field - weighted image space matrix along @xmath19 . each ccd pixel in the focal plane was formed by integrating the focal plane over square regions corresponding to the ccd pixels . the resulting matrix represents the distribution of photons impinging upon each of the ccd elements for each focal plane collected . this matrix is adjusted by a scaling factor to control the signal - to - noise . sub - pixel displacements of the object were produced by shifting the image space grid by one or more of its elements , with respect to the overlying ccd grid . detector noise and background offset were added to each image stack at intensity levels corresponding to experimental values . to represent the photon detection process accurately , poisson noise was simulated for each pixel by random sampling from a poisson distribution with a mean of @xmath1 , where @xmath1 represents the number of photons impinging on each pixel . the intensity units , which represent numbers of photons , were converted to grey scale units by multiplying by the quantum efficiency and by the gain factors of the ccd , and dividing by the photo - electron / adu conversion factor provided in the camera manual @xcite . these values could then be directly converted to integers , that represent grey scale levels in the camera s dynamic range ( 16 bit ) . the signal - to - noise ratio was calculated as the mean pixel value above background of the diffraction - limited spot divided by the standard deviation of the background pixel levels . the mean signal level @xmath100 , mean background level @xmath101 , and variance @xmath102 were calculated in terms of numbers of photons , and the signal - to - noise ratio is expressed as a decibel quantity , @xmath103 the localization error is not a monotonic function of the point - to - point separation . as the displacement decreases below @xmath104 , there is a large increase in the localization error . at some separation below @xmath104 , the error decreases , and then upon further reduction in separation rapidly diverges as the two features become difficult to distinguish . in order to examine this effect more closely , the measured point - to - point separation was compared to the true point - to - point separation . in figure [ fig : img_sim : msep_vs_asep ] . is @xmath105 db.,scaledwidth=80.0% ] we determined that the initial increase in error , at @xmath106 - 9 pixels separation , is due to a systematic over - estimation of the point - to - point separation . the overlap between the intensity distributions forces the least - squares fitting routine to localize the centroid of each intensity distribution away from the overlap region , resulting in overestimation of the point - to - point separation . as overlap increases , the centroids of each feature are forced back towards the overlap region , initially decreasing the estimated separation , until significant overlap has occurred , at @xmath107 pixels , and the two spots are difficult to distinguish . to investigate this behaviour of the localization error at close separations in greater detail , a simulation was performed wherein the features were localized in only one - dimension . three - dimensional images of point features were created as described above . the features were displaced from each other along the @xmath18 direction by a known amount and the @xmath19 coordinate of each point was localized by fitting each intensity distribution independently to a one - dimensional gaussian function . the centroids were displaced at fractional pixel values along the @xmath19 -axis and the integer pixel locations of the centroids were used as initial parameter estimates for the gaussian fits . the @xmath18-direction was chosen because localization of the @xmath19-coordinate has the largest error for closely spaced points . to localize features with limited spatial filtering , i.e. the extreme values of @xmath108 and @xmath109 , the noise had to be reduced to unrealistically small values , mimicking unattainably perfect imaging conditions . for suitable comparison , the simulations in figure [ fig : img_sim:2poinintdist ] were performed with @xmath110 . when filtering is used , it is a source of small overestimation of the separations . filtering removes high spatial frequency components from the image , predominantly noise ; however , the operation also smoothes components of the real intensity distribution that vary rapidly in space . when the distance between points is smaller than @xmath111 , the filtering operation modifies the summed intensity distribution of the two features such that the distance between the two is overestimated by fitting each feature independently to a gaussian function . the origin of the behavior of the localization error for features that are closely spaced is verified upon examining the effect of the band - pass filter . at some separation close to @xmath112 , the effect of filtering in the overlap region becomes negligible compared to the effect of the overlap itself . at this separation the measured separation agrees with the actual separation again ( fig . [ fig : img_sim : msep_vs_asep ] ) . at these and smaller separations , the measured separation begins to systematically diverge from the actual separation , due to overlap of the feature psf s . plots of the intensity distributions before and after filtering for two different spacings are shown in figure [ fig : img_sim:2poinintdist ] . propagation of localization error in the separation of two points was studied in order to determine its role , if any , in the observable @xmath36 , the spindle pole - pole separation . the error for any arbitrary quantity @xmath113 that has been calculated from the set of coordinates for both features @xmath114 can be estimated by propagating the errors associated with each of the coordinates . for notational simplicity , the coordinates representing the positions of each of the features are written as the set @xmath115 , where @xmath116 . the errors for all the coordinates are written generally as @xmath117 , and the covariance terms @xmath118 are given by @xmath119 . the set of values @xmath120 is the set of most probable estimates of the true coordinates , which is exactly known for the simulations . frequently when uncertainties are estimated , the covariant terms are assumed negligible . however , for parameters obtained from fitting a curve to data , the covariant terms can contribute significantly to uncertainties @xcite . the appropriate relation for estimation of uncertainty in @xmath121 is : @xmath122 the error in point - to - point distance can be determined directly during a simulation by comparing the real separation with the measured separation . this error can also be determined from the uncertainties of each coordinate via equation [ eq : errorprop ] , since @xmath123 . to test this , a simulation was performed with the displacement vector between the points oriented in an arbitrary direction in @xmath16 , @xmath17 , and @xmath18 , and the error in @xmath124 determined over a range of separations . the results of this simulation are displayed in figure [ fig : img_sim : errorpropcompare ] . . @xmath125 for this simulation is 10.5 db.,scaledwidth=60.0% ] it can be seen that for large separations , in our case for @xmath126 , the covariant terms are negligible . as the two features approach one another to within less than @xmath127 , the covariant terms become significant and they must be included for an accurate estimate of the error . by simulating point - to - point separation trajectories for each of the directions in the set @xmath128 the form of all @xmath117 , can be obtained . using this relation for @xmath129 , the errors for all quantities determined from feature finding can be estimated . chromatic shift was measured and corrected for prior to combining results obtained using two different wavelengths . the chromatic shift in @xmath2 , @xmath78 , and @xmath19 between the gfp and red acquisition channels was determined by imaging approximately @xmath130 @xmath131-diameter inspeck beads ( invitrogen ) immobilized on a poly - k - coated cover glass . the intensity distribution of each of the beads in a field - of - view was fit to a three - dimensional gaussian function to find the centroid of every bead in each channel . gfp- and red - channel centroid positions were then compared for each of the immobilized beads to determine the chromatic shift between the excitation channels , averaged across all the beads . the @xmath19 position in the 532 nm channel was systematically higher than it was in the 491 nm channel by @xmath132 for the 63x objective lens . the in - plane chromatic shift between the 532 nm and 491 nm channels was @xmath133 and @xmath134 , in @xmath2 and @xmath78 dimensions , respectively . the measured chromatic shift is corrected for in all succeeding analysis where data from the two excitation channels is combined .	we present feature finding and tracking algorithms in 3d in living cells , and demonstrate their utility to measure metrics important in cell biological processes . we developed a computational imaging hybrid approach that combines automated three - dimensional tracking of point - like features , with surface reconstruction from which cell ( or nuclear ) volume , shape , and planes of interest can be extracted . after validation , we applied the technique to real space context - rich dynamics of the mitotic spindle , and cell volume and its relationship to spindle length , in dividing living cells . these methods are additionally useful for automated segregation of pre - anaphase and anaphase spindle populations . we found that genetic deletion of the yeast kinesin-5 mitotic motor cin8 leads to abnormally large mother and daughter cells that were indistinguishable based on size , and that in those cells the spindle length becomes uncorrelated with cell size . the technique can be used to visualize and quantify tracked feature coordinates relative to cell bounding surface landmarks , such as the plane of cell division . by virtue of enriching the information content of subcellular dynamics , these tools will help advance studies of non - equilibrium processes in living cells . locating submicrometer entities at their address in a living cell is an important task in capturing subcellular dynamics . current imaging studies of subcellular , or subnuclear , dynamics of localized features of interest in small bounding volumes , for example in the important model organism budding yeast , the eukaryotic nucleus , or er / golgi bodies , suffer from important limitations . first , dynamics single features such as a chromosome locus position are often described only by position , without contextual information @xcite , subjecting the extracted features to translational , rotational , or vibrational noise , and to lack of knowledge of position in the cell s cycle . features that are free to move in three - dimensional ( 3d ) space are often either detected in two dimensions ( 2d ) @xcite , or in 3d without subpixel resolution in 3d , allowing out - of - plane motion to be ascribed to in - plane motion . in many other studies , contextual information is used , for example relative positions of two spots such as in the spindle pole bodies ( spbs ) , but without subpixel resolution in 3d @xcite , so that the uncertainty in the measurement is insufficiently small to enable new and valuable dynamical information at tens of nanometers scale to be uncovered . there are however important exceptions to these limitations @xcite . finally , in many reports cell volume is extracted by measuring an area and extrapolating to volume by assuming the cell has some regular azimuthal symmetry @xcite , which it may not have . in various micropipette aspiration studies to determine mechanical behavior , the cell has been either considered as a fully incompressible material ( with true 3d volume unmonitored during deformation ) @xcite , or has been assumed compressible without any emphasis on its importance @xcite . yet , a number of cell types have already been shown to be compressible @xcite . cell compressibility may prove to be more general @xcite . on account of these limitations , changes in cell volume due to deformation and other perturbations may be strongly underappreciated . commercial software solutions exist for extracting volume using isosurface methods on a continuous mass distribution and have been applied to cells @xcite . isosurface methods are notoriously sensitive to the threshold chosen . there is a need for methods of reconstructing cell or small subcellular volumes surface that are relatively insensitive to thresholding , and that are robust to challenging conditions of local regions of high curvature . clearly , additional , and more flexible and accessible , high - accuracy computational imaging tools for point and bounded surface features are needed . such methods would not only enable direct volume measurements in combination with feature tracking of localized entities , but would facilitate accurate monitoring of cell volume under nonlinear deformation of cells , and open up new avenues distinguishing between compressibility and viscoelastic bulk relaxation / fluid flow in the nonlinear deformation behavior of cells . here we describe a hybrid computational imaging technique that overcomes these obstacles and generates high resolution tracking of nanoscale to organelle - sized spots in living cells at identifiable locations within the cell , together with bounding surfaces , from hundreds of cells . the watertight bounding surface is obtained using computational geometry techniques and enables monitoring of the volumes of dividing cells and the orientation of the virtual plane separating them . the features can then be transformed to a new co - ordinate system unique to each cell if desired . these methods allow for accurate dissection of yeast mitosis or chromosome locus motions in their precise cellular context _ in vivo_. besides automated high - resolution , high - throughput image and data analysis , our technique also makes possible the extraction of new phenotypic metrics by which cell - to - cell heterogeneity across populations can by quantified @xcite . much cellular motion is stochastic and must be assessed statistically from large populations . the ability to calculate volumes and shapes in situations of low sampling of surface points is likely to become more important as more sophisticated models are developed for how the interior of the cell and nucleus are organized . here , we use these methods to demonstrate nanometer - scale measurements of directed motions or fluctuations of the eukaryotic mitotic spindle , using relative motion to unambiguously identify cells in different stages of cell division ; additionally , to demonstrate measurements of cell volume of individual small biological cells . we show that the resolution of the point feature finding in our optical system , for which we perform simulations of model spindle poles under similar image detection conditions to what is expected in the living cell , is @xmath0 in 3d , depending on the specific signal - to - noise in the acquired images . the volume methods were applied to a large population of non - genetically perturbed cells and populations of cells with mitotic kinesin-5 molecular motors deleted .
a symbolic framework for boundary problems was built up in @xcite for linear ordinary differential equations ( lodes ) ; see also @xcite for more recent developments . one of our long - term goals is to extend this to boundary problems for linear partial differential equations ( lpdes ) . since this is a daunting task in full generality , we want to tackle it in stages of increasing generality . in the first instance , we restrict ourselves to _ constant coefficients _ , where the theory is quite well - developed @xcite . within this class we distinguish the following three stages : 1 . the simplest is the cauchy problem for _ completely reducible _ operators . 2 . the next stage will be the _ cauchy problem _ for general hyperbolic lpdes . 3 . after that we plan to study _ boundary problems _ for elliptic / parabolic lpdes . in this paper we treat the first case ( section [ sec : cauchy - analytic ] ) . but before that we build up a _ general algebraic framework _ ( sections [ sec : boundary - data ] and [ sec : greens - operators ] ) that allows a symbolic description for all boundary problems ( lpdes / lodes , scalar / system , homogeneous / inhomogeneous , elliptic / hyperbolic / parabolic ) . using these concepts and tools we develop a general solution strategy for the cauchy problem in the case ( 1 ) . see the conclusion for some thoughts about the next two steps . the passage from lodes to lpdes was addressed at two earlier occasions : * an _ abstract theory of boundary problems _ was developed in @xcite , including lodes and lpdes as well as linear systems of these . the concepts and results of sections [ sec : boundary - data ] and [ sec : greens - operators ] are built on this foundation , adding crucial concepts whose full scope is only appreciated in the lpde setting : boundary data , semi - homogeneous problem , state operator . * an algebraic language for _ multivariate differential and integral operators _ was introduced in section 4 of @xcite , with a prototype implementation described in section 5 of the same paper . this language is generalized in the pidos algebra of section [ sec : cauchy - analytic ] , and it is also implemented in a mathematica package . in this paper we will not describe the current state of the _ implementation _ ( mainly because of space limitations ) . let us thus say a few words about this here . a complete reimplementation of the pidos package described in @xcite is under way . the new package is called opido ( ordinary and partial integro - differential operators ) , and it is implemented as a standalone mathematica package unlike its predecessor , which was incorporated into the theorema system . in fact , our reimplementation reflects several important design principles of theorema , emphasizing the use of functors and a strong support for modern two - dimensional ( user - controllable ) parsing rules . we have called this programming paradigm funpro , first presented at the mathematica symposium @xcite . the last current stable version of the ( prototype ) package can be found at http://www.kent.ac.uk/smsas/personal/mgr/index.html . at the time of writing , the ring of _ ordinary integro - differential operators _ is completed and the ring of _ partial integro - differential operators _ is close to completion ( for two independent variables ) . compared to @xcite , the new pidos ring contains several crucial new rewrite rules ( instances of the substitution rule for resolving multiple integrals ) . our conjecture is that the new rewrite system is noetherian and confluent but this issue will be analyzed at another occasion . _ notation_. the algebra of @xmath0 matrices over a field @xmath1 is written as @xmath2 , where @xmath3 or @xmath4 is omitted . thus we identify @xmath5 with the space of column vectors and @xmath6 with the space of row vectors . more generally , we have @xmath7 . as mentioned in the introduction , we follow the _ abstract setting _ developed in @xcite . we will motivate and recapitulate some key concepts here , but for a fuller treatment of these issues we must refer the reader to @xcite and its references . let us recall the notion of boundary problem . fix vector spaces @xmath8 and @xmath9 over a common ground field @xmath1 of characteristic zero ( for avoiding trivialities one may assume @xmath8 and @xmath9 to be infinite - dimensional ) . then a _ boundary problem _ @xmath10 consists of an epimorphism @xmath11 and a subspace @xmath12 that is orthogonally closed in the sense defined below . we call @xmath13 the differential operator and @xmath14 the _ boundary space_. similar to the correspondence of ideals / varieties in algebraic geometry , we make use of the following galois connection ( * ? ? ? if @xmath15 is any subspace of the space @xmath8 , its _ orthogonal _ @xmath16 is defined as @xmath17 . dually , for a subspace @xmath14 of the dual space @xmath18 , the orthogonal @xmath19 is defined by @xmath20 . if we think of @xmath8 as `` functions '' and of @xmath18 as `` boundary conditions '' , then @xmath21 is the space of valid conditions ( the boundary conditions satisfied by the given functions ) while @xmath22 is the space of admissible functions ( the functions satisfying the given conditions ) . naturally , a subspace of either @xmath23 of @xmath8 or @xmath18 is called _ orthogonally closed _ if @xmath24 . but while any subspace of @xmath8 itself is always orthogonally closed , this is far from being the case of the subspaces of the dual @xmath18 . hence the condition on boundary spaces @xmath14 to be orthogonally closed is in general not trivial . however , if @xmath14 is finite - dimensional as in boundary problems for lodes ( as in example [ ex : heat - conduction ] below ) , then it is automatically orthogonally closed . for lpdes , the condition of orthogonal closure is important ; see example [ ex : wave - inhom - medium ] for an intuitive explanation . in @xcite and also in the abstract setting of @xcite we have only considered what is sometimes called the _ semi - inhomgeneous boundary problem _ @xcite , more precisely the semi - inhomogeneous incarnation of @xmath25 ; see definition [ def : transfer - operator ] for the full picture . this means we are given a _ forcing function _ @xmath26 and we search for a solution @xmath27 with @xmath28 in other words , @xmath29 satisfies the inhomogeneous `` differential equation '' @xmath30 and the homogeneous `` boundary conditions '' @xmath31 given in @xmath14 . a boundary problem which admits a unique solution @xmath27 for every forcing function @xmath32 is called _ regular_. in terms of the spaces , this condition can be expressed equivalently by requiring that @xmath33 ; see @xcite for further details . in this paper we shall deal exclusively with regular boundary problems . for singular boundary problems we refer the reader to @xcite and @xcite . for a regular boundary problem , one has a linear operator @xmath34 sending @xmath35 to @xmath29 is known as the _ green s operator _ of the boundary problem @xmath25 . from the above we see that @xmath36 is characterized by @xmath37 and @xmath38 . [ ex : heat - conduction ] a classical example of this notion is the two - point boundary problem . as a typical case , consider the simplified model of _ stationary heat conduction _ described by @xmath39 here we can choose @xmath40 for the function space such that the differential operator is given by @xmath41 and the boundary space by the two - dimensional subspace of @xmath42 spanned by the linear functionals @xmath43 and @xmath44 for evaluation on the left and right endpoint . in the sequel we shall write @xmath45 $ ] , employing an important generalization for lpdes , described in the next eamples . we can express its green s operator in the language of integro - differential operators as explained in @xcite . [ ex : wave - inhom - medium ] as a typical counterpart in the world of lpdes , consider the _ equation for waves in an inhomogeneous medium _ , described by @xmath46 in one space dimension . in this case we choose @xmath47 ; again one could choose much larger spaces of functions ( or distributions ) in analysis and in the applications . here the differential operator is @xmath48 while the boundary space @xmath49 is the orthogonal closure of the linear span of the families of functionals @xmath50 and @xmath51 defined by @xmath52 and @xmath53 . using the notation @xmath54 $ ] for denoting the orthogonal closure of the linear span , we can thus write @xmath55 $ ] for the boundary space under consideration . the point of the _ orthogonal closure _ is that the given conditions imply other conditions not in their span , for example @xmath56 or @xmath57 . rather than being linear consequences , these two examples are differential and integral consequences . ( of course the full boundary space also contains many functionals without a natural analytic interpretation . ) in the problems above , the differential equation is inhomogeneous while the boundary conditions are homogeneous . a _ semi - homogeneous boundary problem _ is the opposite , combining a homogeneous differential equation with inhomogeneous boundary conditions . while this is a simple task for lodes ( as always we assume that the fundamental system is available to us in some form ! ) , it is usually a nontrivial problem for lpdes ( even when they have constant coefficients ) . we will give the formal definition of a semi - inhomogeneous boundary problem in the next section ( definition [ def : transfer - operator ] ) . here it suffices to consider an example for developing the necessary auxiliary notions . [ ex : cauchy - wave - eq ] the _ cauchy problem for the wave equation _ in one dimension is @xmath58 being a hyperbolic problem , we could use rather general function spaces for the `` boundary data '' @xmath59 . for reasons of uniformity we will nevertheless restrict ourselves here to the analytic setting , so assume @xmath60 . note that the association of @xmath29 to @xmath61 is again a linear operator mapping ( two univariate ) functions to a ( bivariate ) function ; we will come back to this point in definition [ def : transfer - operator ] . going back to the abstract setting , one is tempted to define the notion of boundary data as some kind of functions depending on `` fewer '' variables . but the problem with this approach is that abstractly speaking we are not dealing with any functions depending on any number of variables ( but see below ) . moreover , the inhomogeneous boundary conditions in the form @xmath62 are basis - dependent while the whole point of the abstract theory is to provide a _ basis - free description _ ( which leads to an elegant setting for describing composition and factorization of abstract boundary problems ) ; see proposition [ prop : bp - product ] . we shall therefore develop a basis - independent notion of boundary data ( we can go back to the traditional description by choosing a basis ) . we define first the _ trace map _ @xmath63 as sending @xmath64 to the functional @xmath65 . in example [ ex : cauchy - wave - eq ] this would map the function @xmath66 to its position and velocity values on @xmath67 . we call @xmath68 the trace of @xmath35 and write it as @xmath69 . moreover , we denote the image of the map @xmath70 by @xmath71 and refer to its elements as _ boundary data_. note that @xmath71 is usually much smaller than the full dual @xmath72 since a continuous function ( let alone an analytic one ) can not assume arbitrary values . ( this situation is vaguely reminiscent of the algebraic and continuous dual of a topological vector space . ) since by definition the trace map is surjective from @xmath8 to @xmath71 , it has some right inverse @xmath73 . we refer to @xmath74 as an _ interpolator _ for @xmath49 since it constructs a `` function '' @xmath75 from given boundary values @xmath76 such that @xmath77 . of course , the choice of @xmath35 is usually far from being unique . apart from its use for describing boundary data ( see at the end of this section ) , the notion of interpolator will turn out to be useful for solving the semi - homogeneous boundary problem ( see proposition [ prop : proj - transfer ] ) . let us now describe how to relate these abstract notions to the usual setting of initial and boundary values problems as they actually in analysis : essentially by _ choosing a basis_. however , we have to be a bit careful since we must deal with the orthogonal closure . [ def : boundary - basis ] if @xmath78 is any orthogonally closed subspace , we call a family @xmath79 a _ boundary basis _ if @xmath80 $ ] , meaning @xmath49 is the orthogonal closure of the span of the @xmath81 . note that a boundary basis is typically smaller than a @xmath1-linear basis of @xmath49 . all traditional boundary problems are given in terms of such a boundary basis . in example [ ex : cauchy - wave - eq ] , the boundary basis could be spelled out by using @xmath82 with @xmath83 and @xmath84 . relative to a boundary basis @xmath79 , we call @xmath85 the _ boundary values _ of @xmath64 . as we can see from the next proposition , we may think of the trace as a basis - free description of boundary values . conversely , one can always extract from any given boundary data @xmath86 the boundary values @xmath87 as its coordinates relative to the boundary basis @xmath88 . [ lem : trc - coord ] let @xmath78 be a boundary space with boundary basis @xmath89 . if for any @xmath90 one has @xmath91 then also @xmath92 . in particular , for any @xmath64 , the trace @xmath69 depends only on the boundary values @xmath93 . since @xmath71 is the image under the trace map , we have @xmath94 and @xmath95 for some @xmath96 . so assume @xmath97 for all @xmath98 and thus @xmath99 for all @xmath100 by the definition of orthogonal closure . then we have @xmath101 and thus @xmath102 . the _ analytic interpretation _ of this proposition is clear in concrete cases like example [ ex : wave - inhom - medium ] : once the values @xmath103 and @xmath104 are fixed , all differential and integral consequences , as in the above examples @xmath56 or @xmath105 , are likewise fixed . it is therefore natural that an interpolator need only consider the boundary values rather than the full trace information . this is the contents of the next lemma . [ lem : int - coord ] let @xmath78 be a boundary space with boundary basis @xmath89 and write @xmath106 for the boundary values of any @xmath64 and @xmath107 for the @xmath1-subspace of @xmath108 generated by all boundary values @xmath109 . then any linear map @xmath110 with @xmath111 induces a unique interpolator @xmath73 defined by @xmath112 . we must show that @xmath73 is a right inverse of @xmath113 . so for arbitrary @xmath114 we must show @xmath115 . by the definition of @xmath71 we can write @xmath94 for some @xmath64 . since @xmath116 , we are left to prove @xmath117 . using lemma [ lem : trc - coord ] , it suffices to prove that @xmath111 , which is true by hypothesis . as noted above , we can always extract the boundary values @xmath118 of some boundary data @xmath86 relative to fixed basis @xmath88 of @xmath49 . however , since one normally has got _ only _ the boundary values ( coming from some function ) , where does the corresponding @xmath119 come from ? by definition , it has to assign values to all @xmath120 , not only to the @xmath81 making up the boundary basis . as suggested by the above lemmata , for actual computations those additional values will be irrelevant . nevertheless , it gives a feeling of confidence to provide these values : if @xmath74 is any _ interpolator _ , we have @xmath121 . this follows immediately from the fact that @xmath74 is a right inverse of the trace map and that it depends only on the boundary values @xmath122 by lemma [ lem : int - coord ] . in the analysis setting this means we interpolate the given boundary value and then do with the resulting function whatever is desired ( like derivatives and integrals in example [ ex : wave - inhom - medium ] ) . using the notion of boundary data developed in the previous section , we can now give the formal definition of the _ semi - homogeneous boundary problem_. in fact , we can distinguish three different incarnations of a `` boundary problem '' ( as we assume regularity , the _ fully homogeneous problem _ is of course trivial ) . [ def : transfer - operator ] let @xmath10 be a regular boundary problem with @xmath11 and boundary space @xmath123 . then we distinguish the following problems : [ cols="^,^,^ " , ] they are , respectively , called the _ fully inhomogeneous _ , the _ semi - inhomogeneous _ and the _ semi - homogeneous _ boundary problem for @xmath10 . the corresponding linear operators will be written as @xmath124 , @xmath125 and @xmath126 , @xmath127 and @xmath128 , @xmath129 . each of the three problems in definition [ def : transfer - operator ] has a unique solution for the respective input data , so the operators @xmath130 are well - defined . employing the usual superposition principle @xmath131 , we can restrict ourselves to the semi - inhomogeneous and the semi - homogeneous boundary problem . for the former , the existence and uniqueness is verified in @xcite . for the semi - homogeneous problem , existence is seen as follows : since the given boundary data @xmath86 can be written as @xmath94 by the definition of @xmath71 , we have to find @xmath132 such that @xmath133 and @xmath134 for all @xmath135 . using the decomposition @xmath136 corresponding to the direct sum @xmath137 , we set @xmath138 . then @xmath133 is clear , and we check further that @xmath139 . for uniqueness , it suffices to prove that @xmath140 has only the trivial solution , which is clear since the sum @xmath141 is direct . the terminology for the operators @xmath130 is not uniform in the literature . in the past , we have only considered @xmath36 and called it the `` green s operator '' acting on a `` forcing function '' @xmath35 . while this is in good keeping with the engineering tradition and large parts of the standard mathematical culture @xcite , it is difficult to combine with suitable terminology for @xmath142 and @xmath143 . in this paper , we shall follow the systems theory jargon @xcite and refer to @xmath142 as the ( full ) _ transfer operator _ , to @xmath36 as the ( zero - state ) signal transfer operator or briefly _ signal operator _ , and to @xmath143 as the ( zero - signal ) state transfer operator or briefly _ state operator_. this terminology reflects the common view of forcing functions @xmath64 as `` signals '' and boundary data @xmath86 as ( initial ) `` states '' . one of the advantages of the abstract formulation is that it allows us to describe the _ product of boundary problems _ in a succinct , basis - free manner ( and it includes lodes and lpdes as well as systems of these ) . the composite boundary problem can then be solved , both in its semi - inhomogneous and its semi - homogeneous incarnation ( the latter is presented here for the first time ) . [ prop : bp - product ] define the product of two boundary problems @xmath10 and @xmath144 with @xmath145 and @xmath146 , @xmath147 by @xmath148 then @xmath149 is regular if both factors are . in that case , if @xmath10 , @xmath150 have , respectively , the signal operators @xmath36 , @xmath151 and the state operators @xmath143 , @xmath152 , then @xmath149 has the signal operator @xmath153 and the state operator @xmath154 acting by @xmath155 . the preservation of regularity and the relation for the signal operators is proved in proposition3.2 of @xcite . for the statement about the composite state operator , note first that the sum @xmath156 is direct by ( * ? ? ? * ( 3.2 ) ) ; hence the definition is consistent . now let @xmath157 be arbitrary boundary data and set @xmath158 . then @xmath159 since @xmath160 and both @xmath161 and @xmath162 vanish by the definition of state operator . hence the differential equation of the composite semi - homogeneous boundary problem is satisfied . it remains to check the boundary conditions @xmath163 for @xmath100 and @xmath164 for @xmath165 . for the first , we use @xmath166 and @xmath167 again to compute @xmath168 . since @xmath169 and @xmath143 is the state operator for @xmath10 , we obtain @xmath170 as required . for the second set of boundary conditions , we use that @xmath171 since @xmath151 is the signal operator with homogeneous boundary conditions @xmath172 . hence @xmath173 , and now the claim follows because @xmath152 is the state operator for the inhomogeneous boundary conditions @xmath165 . as detailed in @xcite , the computation of the _ signal operator _ @xmath36 can be decomposed in two parts : ( 1 ) finding a _ right inverse _ @xmath174 of the differential operator @xmath13 , which involves only the differential equation without boundary conditions ( so we may replace the boundary by intial conditions , thus having again a unique solution : this is the so - called fundamental right inverse ) . ( 2 ) determining the _ projector _ onto the homogeneous solution space along the space of functions admissible for the given boundary conditions the projector `` twists '' the solutions coming from the right inverse into satisfying the boundary conditions . an analogous result holds for the computation of the _ state operator _ @xmath143 if we replace the right inverse @xmath174 of @xmath13 by the interpolator @xmath74 for the boundary space @xmath49 . [ prop : proj - transfer ] let @xmath10 be regular with operators @xmath130 as in definition [ def : transfer - operator ] . then we have @xmath175 and @xmath176 , hence @xmath177 for the transfer operator . here @xmath178 is any right inverse of the differential operator @xmath11 and @xmath73 any interpolator for @xmath49 while @xmath179 is the projector determined by @xmath180 and @xmath181 . the formula for @xmath36 is given in ( * ? ? ? * ( 2.3 ) ) . for proving @xmath176 , let @xmath86 be arbitrary and set @xmath182 . then @xmath133 follows since @xmath183 . furthermore , for every @xmath100 we have @xmath184 by the definition of @xmath74 and @xmath185 . this means that @xmath186 satisfies the boundary conditions , so @xmath143 solves the semi - homogeneous boundary problem for @xmath10 . if @xmath13 is a completely reducible differential operator with constant coefficients in @xmath187 , the determination of @xmath174 reduces to solving an inhomogeneous first - order equation with constant coefficients which is of course straightforward ( lemma [ lem : transport - equation ] ) . also the determination of the interpolator @xmath74 turns out to be easy for a cauchy problem since it is essentially given by the corresponding taylor polynomial . hence it remains to find some means for computing the _ kernel projector _ @xmath188 for a boundary problem @xmath10 . in the case of a lode of order @xmath189 , the method for computing @xmath188 given in the proof of theorem 26 of @xcite and in section 6 of @xcite is essentially a gaussian elimination on the so - called _ evaluation matrix _ @xmath190_{ij } \in k^{n \times n}$ ] formed by evaluating the @xmath191-th boundary condition @xmath81 on the @xmath192-th fundamental solution @xmath193 . so here we assume @xmath194 is a basis of @xmath195 and @xmath196 a basis of @xmath49 . unfortunately , this is not a very intuitive description of @xmath188 , and it is not evident how to generalize it to the lpde case . we have to gain a more conceptual perspective at @xmath197 for making the generalization transparent . let us write @xmath198 for the bilinear operation of evaluation @xmath199 . choosing bases @xmath200 for @xmath49 and @xmath194 for @xmath195 , the coordinate matrix of @xmath201 is clearly @xmath197 . by the usual technique of dualization , we can also think of @xmath201 as the map @xmath202 that sends @xmath203 to the functional @xmath204 . but this map is nothing else than the _ restriction of the trace map _ @xmath113 to @xmath205 . it is easy to check that the restricted trace is bijective and that its inverse gives rise to the projector . [ prop : proj ] let @xmath10 be a regular boundary problem with @xmath206 being the restricted trace map . then @xmath207 is bijective with the state operator @xmath143 as its inverse , and @xmath208 is the projector with @xmath183 and @xmath181 . given any boundary data @xmath86 , we know that @xmath186 satisfies the inhomogeneous boundary conditions @xmath209 so that @xmath210 . hence @xmath211 is a left inverse of @xmath212 . but it is also a right inverse because @xmath213 is a solution of the semi - homogeneous boundary problem with boundary data @xmath214 and thus must coincide with @xmath215 by the uniqueness of solutions . for the projector , observe first that @xmath113 is a left inverse of @xmath128 . as is well known ( * ? ? ? * ( a.16 ) ) , this implies at once that @xmath216 is a projector with @xmath217 and @xmath218 . it remains to show @xmath219 and @xmath220 . the second identity follows from the definition of @xmath70 . for the first identity , the inclusion from left to right follows from the definition of @xmath143 , the reverse inclusion by writing @xmath221 for @xmath222 . we observe that the formula @xmath223 has a very _ natural interpretation _ : the kernel projector picks up the boundary data @xmath224 of an arbitrary function @xmath27 and then constructs the required kernel element @xmath225 by solving the semi - homogeneous boundary problem with boundary data @xmath224 . in the lode case , the relation @xmath208 reduces to the aforementioned formulae ( see @xcite after proposition6.1 ) after choosing bases @xmath194 for @xmath195 and @xmath196 for @xmath49 . apart from its conceptual clarity , the advantage of proposition[prop : proj ] is that it can also be used in the ldpe case ( see after lemma [ lem : transport - equation ] ) . at this point we switch from the abstract setting of sections [ sec : boundary - data ] and [ sec : greens - operators ] to the concrete setting of _ analytic functions_. note that we are dealing with complex - valued functions of real arguments . this means the ground field is @xmath226 , and @xmath8 is the integro - differential algebra of entire functions restricted to real arguments . more precisely , we shall employ the following conventions for easing the burden of book - keeping : as elements of @xmath8 we take all holomorphic functions @xmath227 for any @xmath228 , including the constant functions @xmath229 for @xmath230 . in other words , @xmath8 is a direct limit of algebras . moreover , we have derivations @xmath231 and integrals @xmath232 for all @xmath233 , namely @xmath234 and @xmath235 where @xmath236 occurs at the @xmath189-th position . clearly , we have then integro - differential algebras @xmath237 for every @xmath238 . in fact , @xmath8 has the structure of a _ hierarchical integro - differential algebra_. this notion will be made precise at another occasion ; for the moment it suffices to make the following observations . if @xmath239 is the sublattice of the powerset @xmath240 that consists of finite sets @xmath241 and the full set @xmath242 , we define for @xmath243 the subalgebras @xmath244 consisting of the functions depending ( at most ) on @xmath245 . then @xmath246 is a sublattice of @xmath247 that is isomorphic to the lattice @xmath248 . the bottom element is of course @xmath249 , the top element @xmath250 . we write @xmath251 as an abbreviation for @xmath252 . as in the earlier paper @xcite , we add to this algebraic structure all linear _ substitution operators_. in accordance with the above hierarchical structure , we use the ring @xmath253 of row and column finite matrices with complex entries . is of course also analytic on @xmath254 with values in @xmath187 , so there is no problem with this view . for example , the substitution @xmath255 sends @xmath256 to @xmath257 . moreover , complex substitutions are indispensible for specifying the general solution of elliptic equations like the laplace equation . ] this means any @xmath258 can actually be seen as a finite matrix @xmath259 with @xmath260 rows and @xmath189 columns , extended by zero rows and columns . as usual , we identify @xmath261 with the linear map @xmath262 , yielding the substitution operator @xmath263 defined by @xmath264 . we write @xmath265 $ ] for the _ pidos algebra _ generated over @xmath187 by the operators @xmath266 , the substitutions @xmath267 induced by @xmath268 and the exponential basis polynomials @xmath269 . here @xmath270 denotes the arguments @xmath271 for any @xmath272 , with exponents @xmath273 and frequencies @xmath274 . obviously , @xmath265 $ ] acts on @xmath8 , with @xmath275 acting as @xmath276 and @xmath277 as @xmath278 , similar to the univariate case in the older notation of @xcite . here we avoid the notation @xmath279 for the integrals since the powers @xmath280 might be mistaken as integrals with upper bound @xmath189 . the algebra @xmath265 $ ] can be described by a _ rewrite system _ ( pidos = partial integro - differential operator system ) , analogous to the one given in @xcite . we will present this system in more detail in particular proofs of termination and confluence at another occasion . since in this paper we restrict ourselves to the analytic setting , we can appeal to the well - known _ cauchy - kovalevskaya theorem _ * thm . 2.22 ) for ensuring the existence and uniqueness of the solution of the cauchy problem . while the theorem in its usual form yields only local results , there is also a global version ( * ? ? ? 7.4 ) that provides a good foundation for our current purposes . since this form of the theorem is not widely known , we repeat the statement here . as usual , we designate one _ lead variable _ @xmath281 , writing the other ones @xmath282 as before . note that in applications @xmath281 is not necessarily time . the apparently special form of the differential equation @xmath133 implies no loss of generality : whenever @xmath283 $ ] is a differential operator of order @xmath260 , the change of variables @xmath284 leads to an equation of the required form . [ thm : global - kovalevskaya ] let @xmath285 $ ] be a differential operator in caucy - kovalevskaya form with respect to @xmath281 , meaning @xmath286 with @xmath287 and @xmath288 . then the cauchy problem @xmath289 has a unique solution @xmath290 for given @xmath291 . in the _ abstract language _ of sections [ sec : boundary - data ] and [ sec : greens - operators ] this is the semi - homogeneous boundary problem @xmath10 with boundary space @xmath292,\ ] ] where the evaluation @xmath293 is written as the substitution @xmath294 denotes . hence the solution of is given by the state operator @xmath295 if we identify the boundary data @xmath86 with its coordinate representation @xmath296 relative to the above boundary basis @xmath297 . in detail , @xmath298 is the unique linear map sending @xmath299 to @xmath300 ; confer lemma [ lem : trc - coord ] for the uniqueness statement . in the sequel these identifications will be implicit . for future reference , we mention also that the usual taylor polynomial allows one to provide a natural _ interpolator _ for the initial data , namely @xmath301 which we will not need here because compute the kernel projector directly from its first - order factors . in this paper , we will study the cauchy problem for a _ completely reducible operator _ @xmath302 , meaning one whose characteristic polynomial @xmath303 = { \mathbb{c}}[\lambda_1 , \dots , \lambda_n]$ ] splits into linear factors . hence assume @xmath304 with first - order operators @xmath305 $ ] . by a well - known consequence of the ehrenpreis - palamodov theorem , the general solution of @xmath133 is the sum of the general solutions of the factor equations @xmath306 ; see the corollary on @xcite . hence it remains to consider differential operators that are powers of first - order ones ( we may assume all nonconstant coefficients are nonzero since otherwise we reduce @xmath189 after renaming variables ) . [ lem : power - first - order ] let @xmath307 $ ] be a first - order operator with all @xmath308 . order the variables such that all cumulative sums @xmath309 are nonzero . then the general solution of @xmath310 is given by @xmath311 where @xmath312 are arbitrary functions of the indicated arguments . this can be found in some textbooks on differential equations @xcite . setting up the change of variables given by and @xmath313 , the differential operator is @xmath314 in the new coordinates ( the ordering ensures invertibility ) . clearly , its fundamental solutions are @xmath315 , and the `` integration constants '' @xmath316 are arbitrary functions of @xmath317 . in principle , one could now combine the general solutions for each factor , substitute them into the initial conditions of and then solve for the @xmath316 in terms of the prescribed boundary data @xmath318 . with this choice of @xmath316 , the general solution will become the state operator for the cauchy problem . however , this is a very laborious procedure , and therefore we prefer to use another route . since we assume a completely reducible operator , we can employ the product representation of proposition [ prop : bp - product ] . in that case , it remains to consider the case of a _ single first - order factor_. [ lem : transport - equation ] let @xmath307 $ ] be a first - order operator with all @xmath308 . then the cauchy problem @xmath133 , @xmath319 has the state operator @xmath320 . moreover , the differential operator @xmath13 has the right inverse @xmath321 here @xmath322 is the transformation with @xmath313 , and @xmath323 is its inverse . this follows immediately from lemma [ lem : power - first - order ] . the right inverse is computed using lemma 3 of @xcite after transforming the lpde to a lode . by proposition[prop : proj ] , we can determine the _ kernel projector _ for the cauchy problem of lemma [ lem : transport - equation ] as @xmath208 , where @xmath324 in this simple case . having the kernel projector and the right inverse @xmath174 in lemma [ lem : transport - equation ] , the _ signal operator _ is computed by @xmath325 as usual . now we can tackle the general cauchy problem by a simple special case of proposition [ prop : bp - product ] . [ prop : ivp - product ] let @xmath326 $ ] be two first - order operators with nonzero coefficients for @xmath327 . if @xmath328 is the evaluation defined after theorem . [ thm : global - kovalevskaya ] , then we have @xmath329 ) \ , ( t_2 , [ l_{0 , \xi } \mid \xi \in { \mathbb{r } } ] ) = ( t_1 t_2 , [ l_{0 , \xi } , l_{0 , \xi } d_t \mid \xi \in { \mathbb{r}}]\ ] ] for the product of the cauchy problems . by the definition of the product of boundary problems , we have to show that @xmath330 = [ l_{0 , \xi } , l_{0 , \xi } t_2]$ ] . since each of these is defined as a biorthogonal , it suffices to prove that the system @xmath331 has the same solutions as the system @xmath332 . but the latter is given by @xmath333 where the first term and the @xmath334-derivatives vanish since @xmath335 . using @xmath336 , this implies that the two systems are indeed equivalent . this settles the completely reducible case : using proposition [ prop : ivp - product ] we can break down the _ general cauchy problem _ into first - order factors with single initial conditions . for each of these we compute the state and signal operator via lemma [ lem : transport - equation ] , hence the state and signal operator of by proposition [ prop : bp - product ] . as explained in the introduction , we see the framework developed in this paper as the first stage of a more ambitious endeavor aimed at boundary problems for general constant - coefficient ( and other ) lpdes . following the enumeration of the introduction , the next steps are as follows : 1 . stage ( 1 ) was presented in this paper , but the _ detailed implementation _ for some of the methods explained here is still ongoing . the crucial feature of this stage is that it allows us to stay within the ( rather narrow ) confines of the pidos algebra . in particular , no fourier transformations are needed in this case , so the analytic setting is entirely sufficient . 2 . as we enter stage ( 2 ) , it appears to be necessary to employ stronger tools . the most popular choice is certainly the _ framework of fourier transforms _ ( and the related laplace transforms ) . while this can be algebraized in a manner completely analogous to the pidos algebra , the issue of choosing the right function space becomes more pressing : clearly one has to leave the holomorphic setting for more analysis - flavoured spaces like the schwartz class or functions with compact support . ( as of now we stop short of using distributions since that would necessitate a more radical departure , forcing us to give up rings in favor of modules . ) 3 . for the treatment of genuine boundary problems in stage ( 3 ) our plan is to use a powerful generalization of the fourier transformation the _ ehrenpreis - palamodov integral representation _ @xcite , also applicable to systems of lpdes . much of this is still far away . but the _ general algebraic framework _ for boundary problems from sections [ sec : boundary - data ] and [ sec : greens - operators ] is applicable , so the main work ahead of us is to identify reasonable classes of lpdes and boundary problems that admit a symbolic treatment of one sort or another . anja korporal , georg regensburger , and markus rosenkranz . regular and singular boundary problems in maple . in _ proceedings of the 13th international workshop on computer algebra in scientific computing , casc2011 ( kassel , germany , september 5 - 9 , 2011 ) _ , volume 6885 of _ lecture notes in computer science_. springer , 2011 . anja korporal , georg regensburger , and markus rosenkranz . symbolic computation for ordinary boundary problems in maple . in _ proceedings of the 37th international symposium on symbolic and algebraic computation ( issac12 ) software presentation . ulrich oberst and franz pauer . the constructive solution of linear systems of partial difference and differential equations with constant coefficients . , 12(3 - 4):253308 , 2001 . special issue : applications of grbner bases to multidimensional systems and signal processing . markus rosenkranz , georg regensburger , loredana tec , and bruno buchberger . a symbolic framework for operations on linear boundary problems . in vladimir p. gerdt , ernst w. mayr , and evgenii h. vorozhtsov , editors , _ computer algebra in scientific computing . proceedings of the 11th international workshop ( casc 2009 ) _ , volume 5743 of _ lncs _ , pages 269283 , berlin , 2009 . springer . markus rosenkranz , georg regensburger , loredana tec , and bruno buchberger . symbolic analysis of boundary problems : from rewriting to parametrized grbner bases . in ulrich langer and peter paule , editors , _ numerical and symbolic scientific computing : progress and prospects _ , pages 273331 . springer , 2012 .	we introduce a general algebraic setting for describing linear boundary problems in a symbolic computation context , with emphasis on the case of partial differential equations . the general setting is then applied to the cauchy problem for completely reducible partial differential equations with constant coefficients . while we concentrate on the theoretical features in this paper , the underlying operator ring is implemented and provides a sufficient basis for all methods presented here .
excitons in semiconductors have been promising candidates for the observation of bose einstein condensation for several decades . at present , cuprous oxide ( cu@xmath1o ) is in the focus of experimental efforts due to the large binding energy and long lifetime of the exciton states . in order to obtain sufficiently high densities , entrapment by an external potential is an approved method . the theoretical description of excitons in potential traps has been carried out so far mostly in the frame of a model of ideal bosons . concepts for the inclusion of the interaction are well known from the theory of atomic condensates , @xcite and first applications to excitons exist , too . @xcite recent investigations in the framework of a mean - field formalism in local density approximation have shown distinct signatures of a condensate in the decay luminescence spectrum of the non - condensed excitons . @xcite it is the aim of the present paper to introduce a generalization of this theory to a multi - component gas of interacting paraexcitons and orthoexcitons , where the consequences of the interaction on the condensation process are of particular interest . we show results for the densities of the individual components and their spatially resolved luminescence spectra for several parameter regimes and highlight experimentally relevant cases . the thermodynamics of a one - component bose gas has been investigated in detail , see , e.g. , refs . . first applications of these concepts to excitons have been presented in ref . and , looking at spectral signatures of a condensate , in ref . . in analogy to generalizations for multi - component atomic gases , e.g. , refs . and spinor polaritons , e.g. , refs . , in the following , we generalize this approach to the case of multiple species of excitons , i.e. , paraexcitons and orthoexcitons , adopting a mean - field coupling scheme between the components . @xcite the multi - component exciton gas is considered in second quantization . we start from the hamiltonian for a @xmath2-component system in the grand canonical ensemble : @xmath3 with respective external potentials @xmath4 and chemical potentials @xmath5 for each species . we assume a contact potential for the exciton exciton interaction , with the matrix @xmath6 containing the intra- and inter - species interaction strengths . its components are given by the s - wave scattering lengths @xmath7 : @xmath8 the bose field operator @xmath9 obeys the heisenberg equation of motion @xmath10 we decompose the field operators @xmath9 in the usual fashion , @xmath11 where @xmath12 is the ( scalar ) condensate wave function with @xmath13 and @xmath14 is the operator of the thermal excitons . inserting the decomposition ( [ eq : aufteilung ] ) into eq . ( [ eq : heisenberg ] ) and following the steps of ref . , we obtain ( arguments dropped for the sake of brevity ) @xmath15 and @xmath16 with @xmath17 , @xmath18 , and the normal and anomalous averages @xmath19 and @xmath20 , respectively . equation ( [ eq : gpe3kallgerg ] ) generalizes the familiar gross pitaevskii equation ( gpe ) by including ( i ) the coupling to the thermal excitons and ( ii ) the coupling of multiple components . in a first approximation , we neglect all non - diagonal averages , i.e. , @xmath21 , reducing ( [ eq : gpe3kallgerg ] ) and ( [ eq : beweg3kallgerg ] ) to effective one - species equations with an additional mean field contribution from the other species : @xmath22 @xmath23 thanks to this simplification , eq . ( [ eq : beweg3kallgerghfb ] ) can be formally solved by a bogoliubov transformation , @xmath24\,,\ ] ] where @xmath25 enumerates the quasiparticle states . thereby the bogoliubov amplitudes @xmath26 and @xmath27 satisfy the relation @xmath28=1 $ ] . the excitation spectrum @xmath29 is given by the solution of the eigenvalue problem @xmath30 with @xmath31 equations ( [ eq : gpe3kallgerghfb ] ) and ( [ eq : bogomatrix ] ) are the multi - component generalizations of the hartree fock bogoliubov ( hfb ) equations . they represent a system of @xmath32 equations which are coupled via the interaction matrix elements @xmath6 . to guarantee gapless spectra , we next apply the popov approximation , i.e. , we neglect the anomalous averages @xmath33 in ( [ eq : gpe3kallgerghfb ] ) and ( [ eq : bogomatrix ] ) and find @xmath34 and @xmath35 since the extension of the potential trap is large compared to the typical length scale of the system ( e.g. , the thermal debroglie wavelength of the excitons ) , we can use the local density approximation ( lda ) . then the excitons are treated as a locally homogeneous system and the spatial dependence enters only via the trap potential . in that case , the bogoliubov equations are readily solved , yielding the density @xmath36 of thermally excited excitons as @xmath37\nonumber\\ & & \times\theta\left(e_i(\mathbf{k},\mathbf{r})^2\right ) \end{aligned}\ ] ] with @xmath38^{-1}$ ] being the usual bose function . the excitation spectrum @xmath39 is explicitly given by @xmath40 with @xmath41 and @xmath42 . in consistence with the lda , we apply the thomas fermi approximation to the gross pitaevskii equation , neglecting the kinetic energy term in ( [ eq : gpe3kallgerghfbp ] ) . then we obtain finally for the densities of the condensates @xmath43 the expressions ( [ eq : nt3k])([eq : nc3k ] ) have to be solved self - consistently . although they look very similar to the one - component case , a coupling between the components appears via @xmath44 and @xmath45 . in what follows we calculate the densities of excitons in cu@xmath1o in a strain induced potential trap . @xcite in addition to the paraexcitons ( labeled ` p ' ) , two spin projections of orthoexcitons are captured by the trap , denoted by ` + ' and ` @xmath46 ' , while the zero component is expelled and plays no role . thus , the number of components @xmath47 . in addition to the usual symmetry of the interaction matrix , @xmath48 with @xmath49 , in our case it holds that @xmath50 and @xmath51 , leaving four independent parameters @xmath52 , @xmath53 , @xmath54 , and @xmath55 . as extensive works on two - component systems @xcite have shown , one of the most interesting aspects of multicomponent systems the occurrence of phase separation is closely tied to the proportions of inter- and intra - species interaction strengths . according to eq . ( [ eq : hij ] ) , the interaction strengths are given by the s - wave scattering lengths of the corresponding channels , which can be obtained by the solution of the four - particle scattering problem . the case of positronium positronium scattering some time ago received much attention @xcite and quite reliable values of the scattering length for both the singlet and the triplet channel have been obtained . in contrast , the description of exciton exciton interaction is a long - standing problem , and so far no satisfying solution for the general case has been obtained . especially for cu@xmath1o , we expect a strong effect of the non - parabolicity of the valence band @xcite and of the rather large electron hole exchange interaction @xcite on the scattering lengths . therefore , the values we deduced from the scattering lengths of the positronium problem given in ref . ( @xmath56 , @xmath57 and @xmath58 with @xmath59 ev m@xmath60 ) , should be considered as representative , only . nevertheless , they allow to show the general behaviour of the multi - component exciton system . in the calculation we also neglect the difference in the paraexciton and orthoexciton mass due to the @xmath61-dependent exchange interaction . excitons decay by emitting photons . this takes place either directly , whereby momentum conservation requires that only excitons with the same momentum as the emitted photons are involved , or with assistance of momentum supplying phonons such that all exciton states can participate in the optical emission . because the optical wavelength of the emission is much smaller than the trapped exciton cloud , we apply a local approximation to the emission spectrum , which , for the homogeneous case , is determined by the excitonic spectral function @xmath62 @xcite : @xmath63 with @xmath64 representing the exciton - photon coupling . the spectral function is given by the bogoliubov amplitudes @xmath26 and @xmath27 and the quasiparticle spectrum in ( [ eq : energie3k ] ) : @xmath65 \,.\nonumber\end{aligned}\ ] ] in order to account for a finite spectral resolution being important for comparison with measured spectra , we convolute the spectral intensity ( [ eqn : spectrum ] ) with a slit function of the shape exp@xmath66 $ ] . here , @xmath67 is a measure for the spectral resolution . furthermore , in a typical experimental situation , one images a small stripe of width @xmath68 elongated along the @xmath69-direction onto the entrance slit of a spectrograph . integrating over the @xmath70-direction perpendicular to @xmath69 we obtain the spatially resolved spectrum of the thermal excitons ( in the following , the direct condensate contribution is not considered ) : @xmath71 } \nonumber \\ & & \hspace{-0.5 cm } -\int \limits_{-\delta x}^{\delta x } { \mathrm{d}}x \int \limits_{-\infty}^{\infty } { \mathrm{d}}y \int { \mathrm{d}}{\mathbf{k } } \ , \|s_i(\mathbf{k})\|^2\,v_i^2(\mathbf{k},\mathbf{r } ) \nonumber \\ & & \hspace{2ex}\times n_b(-e_i(\mathbf{k},\mathbf{r}))\exp { [ -\varepsilon_{+}^4 ( \omega ' , \mathbf{k } , \mathbf{r})]}\end{aligned}\ ] ] with @xmath72 . in the case of phonon - assisted transitions ( being relevant for the orthoexcitons ) , we have @xmath73 with @xmath74 being the excitonic band gap of the semiconductor . we assume @xmath75 to be a constant . then the first term in eq . ( [ eqn : spectrum ] ) gives rise to a @xmath76-shaped luminescence line at the position of the chemical potential , with a strength determined by the condensate density . for trapped paraexcitons , the zero - phonon decay is relevant and can be treated by setting @xmath77 and @xmath78 . here @xmath79 is the wavevector of the intersection of photon and exciton dispersion . its modulus is given by @xmath80 , where @xmath81 is the refraction index and @xmath82 is the vacuum velocity of light . due to the form of @xmath83 , the condensate itself does not contribute to the direct luminescence process . however , as discussed for the one - component exciton gas in ref . , in case of a potential trap there will be indirect signatures of the condensate in the spatially resolved luminescence spectrum . the spectral lineshape follows the density distribution in the trap , which in turn is bordered by the minimal excitation energy @xmath84 . for a noncondensed gas the latter quantity is roughly parabolic due to the trapping potential , while it is zero in the presence of a condensate . thus with increasing particle number ( or decreasing temperature ) the flat bottom of the spectrum of thermal excitons may be a footprint of bose - einstein condensation . we evaluate the density distributions of the trapped excitons in an iterative way . in each step , we keep the distributions of two of the components fixed . under the constraint of fixed particle number , we iterate the subset of equations ( [ eq : nt3k])([eq : nc3k ] ) belonging to the third component to self - consistency . we cycle through the components until self - consistency of the whole system ( [ eq : nt3k])([eq : nc3k ] ) is achieved . depending on the temperature and particle numbers in the trap , there may occur six distinct situations , featuring a condensate of ( i ) none of the species , ( ii ) only the paraexcitons , ( iii ) only one species of orthoexcitons , ( iv ) both species of orthoexcitons , ( v ) paraexcitons and one species of orthoexcitons , and ( vi ) all species . to analyze these cases , we set the particle numbers of each component to one of two values : @xmath85 or @xmath86 for @xmath87 , respectively . we get a rough estimate of the corresponding critical temperatures by applying a harmonic approximation to the hertzian potentials . then , a simple thomas - fermi calculation for the single component case @xcite yields @xmath88 for @xmath89 and @xmath90 for @xmath91 . for our calculations , we use values of @xmath92 ev for the spectral resolution and @xmath93 m for the entrance slit of the spectrograph being typical for a triple high resolution spectrograph used in the current experiments which are under way . @xcite in the following figures we show the respective trap potentials @xmath4 , minimal excitation energies @xmath94 , and density distributions of paraexcitons and orthoexcitons versus the @xmath69-coordinate . we compare the results to the case without inter - component interaction ( labeled ` i d ' ) . we first investigate case ( i ) by setting @xmath95 , @xmath96 and a temperature @xmath97 k well above both of the estimated critical values . as fig . [ fig1 ] shows , in the absence of any condensate , the lineshapes @xmath94 roughly follow the external potentials and the presence of multiple components causes only a weak additional renormalization . both para- and orthoexciton densities the latter being equal for ` + ' and ` @xmath46 ' species concentrate in the centers of their traps . a noticeable redistribution of the orthoexcitons with respect to the one - component case results from the large number of paraexcitons as well as from the ortho - ortho interaction . the corresponding luminescence spectra of thermal excitons are shown in fig . [ fig1 ] , lower row . because the modulus of the photon vector @xmath98 m@xmath99 is rather small , the integrated zero - phonon spectrum of the paraexcitons almost directly resembles the minimal excitation energy @xmath100 . in the case of orthoexcitons , every @xmath101-vector contributes and we find a broad energy distribution above @xmath102 . keeping the particle numbers constant , we lower the temperature to @xmath103 k and show case ( ii ) in fig . [ fig3 ] . now the renormalized potential of the paraexcitons ( fig . [ fig3 ] , upper left panel ) is cut at the chemical potential causing an almost flat bottom of the luminescence spectrum ( fig . [ fig3 ] , lower left panel ) . again the densities of thermal and condensed paraexcitons show no significant deviation from the single component case ( fig . [ fig3 ] , middle left panel ) . in contrast , while isolated orthoexcitons would have been condensed , there is no ortho - condensate in the fully interacting case , which shows that the presence of multiple repulsive components lowers the critical temperature . due to the even higher concentration of paraexcitons and the different minimum positions of the external potentials ( @xmath104 m and @xmath105 m ) , the thermal orthoexcitons are slightly pushed aside ( fig . [ fig3 ] , middle right panel ) . their spectrum ( fig . [ fig3 ] , lower right panel ) is qualitatively nearly unchanged with respect to case ( i ) . however , due to the lower temperature , the spectrum is less widespread . furthermore , the chemical potential nearly touches the renormalized potential causing already a smoother curvature of the spectral shape . if we exchange the particle numbers of paraexcitons and one species of orthoexcitons , i.e. @xmath106 , @xmath107 , we realize case ( iii ) , which is presented in fig . while in this case the density distributions of thermal and condensed ortho(+)excitons deviate only weakly from the one - component case , the displacement of the thermal paraexcitons is expressed in a heavily distorted zero - phonon spectrum . increasing also the particle number of the remaining ortho - species by an order of magnitude , we generate case ( iv ) , depicted in fig . [ fig7 ] . as shi _ _ et al.__@xcite showed , even for finite temperature the condition for phase - separation of mutually interacting trapped condensates coincides with the @xmath108 result of ho and shenoy : @xcite @xmath109 . due to their strong repulsion , the two ortho - condensates fulfill this condition and separate into a ball - and - shell structure with finite overlap , as seen in refs . and . yet , as found in ref . , at @xmath110 no pure ` + ' - or ` @xmath46'-phases exist and the respective thermal particles are not entirely expelled . references and pointed out that in general the component with the weaker self - interaction forms the outer shell . in the case of orthoexcitons , this criterion does not apply and the labels ` + ' and ` @xmath46 ' can be interchanged in all the results presented here . in principle , there should exist an unstable solution with equal distributions of the ortho - condensates . the breaking of this symmetry is a consequence of our iterative numerical method . because of the symmetry of the interaction , the paraexcitons react to the combined density of the orthoexciton species . that is why the distortion of the para - spectrum is strongest in the area of overlapping ortho - condensates . let us switch back to the parameters of case ( ii ) , i.e. , @xmath111 and @xmath112 , and lower the temperature to @xmath113 k. as fig . [ fig9 ] shows , besides condensed paraexcitons we now find a small condensate of only one species of orthoexcitons ( case v ) . while in the single - component case , for @xmath114 , both ortho - species have the same critical temperature , now the mutual repulsion prevents the simultaneous condensation of the second species . we have to lower the temperature to @xmath115 k , to get condensates of all the components ( case ( vi ) , fig . [ fig11 ] ) . while the ortho - species again form a ball - and - shell structure , ortho- and para - condensates do not separate because of their weak interaction ( @xmath116 ) . we have presented a theoretical approach for the description of multi - component interacting excitonic gases in potential traps . the resulting system of equations has been subject to a number of approximations to make it numerically feasible . finally , coupled multi - component equations for the densities of thermal excitons in hartree fock bogoliubov popov approximation and the condensate densities following from the gross pitaevskii equation in thomas fermi approximation have been obtained and numerically solved . compared to previous calculations , @xcite an experimentally realistic , anharmonic trap potential has been used . six `` typical '' ( but not necessarily experimentally realizable ) situations leading to bec in one or more of the species have been compared . the spatially resolved decay luminescence spectra of thermal paraexcitons and orthoexcitons exhibit clear signatures of a condensate . on the one hand , there is a flat bottom at the chemical potential known from the single - component case . @xcite on the other hand , the interspecies interaction causes additional modifications of the spectra : if there is a condensate in one of the species , the spectrum of the respective other component is clearly distorted . in a typical experiment in the bulk or involving rather shallow potential traps , orthoexcitons are produced by laser excitation , but are converted fast into paraexcitons at a rate of 0.3 ns@xmath99 . @xcite therefore , under quasiequilibrium conditions , the particle number ratio ortho-/paraexcitons is small which corresponds to the cases ( i ) , ( ii ) , ( v ) , and ( vi ) . by increasing the stress , however , the conversion rate decreases by more than an order of magnitude . @xcite by continuous excitation of orthoexcitons , therefore , it should be possible to obtain higher orthoexciton / paraexciton ratios like in cases ( iii ) and ( iv ) . for an exciton number of 5@xmath117 , the density in the center of the trap is about @xmath118 @xmath119 which is experimentally achievable . the same holds for the considered temperatures of @xmath120 k. @xcite table [ table1 ] summarizes the essential information obtained from the cases ( i)(vi ) discussed above . obviously , three conclusions can be drawn : first , if the temperature is low enough ( below the respective critical temperature ) , every species can form a condensate . its primary signature is a flat bottom of the respective spectrum . second , a spatial separation occurs only between the condensates of the two ortho species , because of their strong repulsion . it shows up only in the densities , not in the combined spectrum . therefore , it is important for future experiments to measure also the spectrally integrated density profile . third , at occurrence of any ortho but no para condensate , the para spectrum is distorted in a characteristic way indicating a condensate in at least one of the _ other _ species . the presented theory is obviously only a first step towards a deeper understanding of the physics of trapped excitons . to go beyond that includes the solution of the gross pitaevskii equation ( without the thomas fermi approximation ) and the inclusion of anomalous densities already on the single - component level . moreover , a general multi - component theory requires the consideration of mixed averages which overcomes the effective single - component picture but complicates the bogoliubov transformation remarkably . 8\|c\| case & temperature & & & & & orthoexciton & paraspectrum + no . & ( k ) & @xmath121 & @xmath122 & @xmath123 & condensate ? & separation ? & deformation ? + ( i ) & 2.2 & 1 & 0.1 & 0.1 & - & - & - + ( ii ) & 1.2 & 1 & 0.1 & 0.1 & @xmath124 & - & - + ( iii ) & 1.2 & 0.1 & 0.1 & 1 & @xmath125 & - & + ( iv ) & 1.2 & 0.1 & 1 & 1 & @xmath126 & & + ( v ) & 1.05 & 1 & 0.1 & 0.1 & @xmath127 & - & - + ( vi ) & 0.8 & 1 & 0.1 & 0.1 & @xmath128 & & - + we would like to thank g. manzke and f. richter ( rostock ) , and a. alvermann ( greifswald ) for many fruitful discussions . this work was supported by the deutsche forschungsgemeinschaft via collaborative research center sfb 652 , projects b1 and b5 .	the system under consideration is a multi - component gas of interacting paraexcitons and orthoexcitons confined in a three dimensional potential trap . we calculate the spatially resolved optical emission spectrum due to interband transitions involving weak direct and phonon mediated exciton - photon interactions . for each component , the occurrence of a bose - einstein condensate changes the spectrum in a characteristic way so that it directly reflects the constant chemical potential of the excitons and the renormalization of the quasiparticle excitation spectrum . moreover , the interaction between the components leads , in dependence on temperature and particle number , to modifications of the spectra indicating phase separation of the subsystems . typical examples of density profiles and luminescence spectra of ground - state paraexcitons and orthoexcitons in @xmath0 are given .
observations at all wavelengths from radio to x - rays are consistent with a scenario that tidal interactions between m82 and another member galaxy of the m81 group has channeled large amounts of gas into the central regions of m82 during the last several 100 myr ( e.g. , telesco 1988 , rieke et al . 1993 ) . this has induced a starburst which has continued for up to about 50 myr . all of the bright radio and infrared sources associated with the active starburst are confined to the galaxy s center , lying within a radius of @xmath3 pc , and corresponding spatially with bright optical structures , labeled m82 a , c , and e in oconnell & mangano ( 1978 , om78 ) . however , evidence exists that this is not the only major starburst episode to have occurred in m82 . a region at about 1 kpc ne from the galactic center , m82 b ( cf . om78 ) , has exactly the properties one would predict for a fossil starburst with a similar amplitude to the active burst . marcum & oconnell ( 1996 ) find a sharp main - sequence cut - off of the composite stellar system in m82 b , corresponding to an age of @xmath4 100200 myr and an average extinction of @xmath5 mag . region a , on the other hand , is only consistent with a very young population ( @xmath6 myr ) and is more heavily affected by extinction ( @xmath7 mag ) . by extrapolating region b s surface brightness ( @xmath8 mag arcsec@xmath9 [ om78 ] , after correction for foreground extinction ) back to an age of 10 myr we estimate that its surface brightness was @xmath10 magnitudes brighter ( cf . the bruzual & charlot [ 1996 , bc96 ] stellar population models ) , similar to that presently observed in the active starburst . _ hst _ imaging of the bright central regions of m82 resolved these into a swarm of young star cluster candidates , with a fwhm of @xmath11 pc ( @xmath12 ) and mean @xmath13 ( oconnell et al . 1994 , 1995 ) , brighter than any globular cluster in the local group . this is the nearest rich system of such objects ; such `` super star clusters '' have also been discovered with _ hst _ in other interacting and amorphous systems , and in dwarf and starburst galaxies ( e.g. , holtzman et al . 1992 , whitmore et al . 1993 , oconnell et al . 1994 , conti et al . 1996 , ho 1997 , carlson et al . 1998 , watson et al . 1998 , among others ) . their diameters , luminosities , and in several cases masses are consistent with these being young _ globular _ clusters formed as a result of recent gas flows ( e.g. , van den bergh 1995 , meurer 1995 , ho & filippenko 1996 ) . it is possible that most of the star formation in starbursts takes place in the form of such concentrated clusters . our observations of m82 do not reveal similar cluster formation outside the active and the fossil starburst regions . under the assumption that region b is indeed a fossil starburst site , it is expected that it originally contained a complement of luminous clusters similar to that now observed in region a. the combination of observations of both the active and the fossil starburst sites in m82 therefore provides a unique physical environment for the study of the stellar and dynamical evolution of these star cluster systems . the fossil starburst region , m82 b , was observed on september 15 , 1997 , with both _ wfpc2 _ and _ nicmos _ on board the _ hst_. we imaged two adjacent @xmath14 square fields ( planetary camera [ pc ] field of view , @xmath15 pix@xmath16 ) in the m82 b region in the f439w , f555w and f814w passbands , with total integration times of 4400s , 2500s and 2200s , respectively , for region `` b1 '' , and 4100s , 3100s and 2200s , respectively , for region `` b2 '' . these observations were obtained using four exposures per filter , covering a large range in integration times to facilitate the removal of cosmic ray events . the f439w , f555w and f814w filters have roughly similar characteristics to the johnson - cousins broad - band _ b , v _ and _ i _ filters , respectively . in the near - infrared ( nir ) we chose to use _ nicmos _ camera-2 ( @xmath17 pix@xmath16 ) , which provided the best compromise of resolution and field of view . we acquired 4 partially spatially overlapping exposures in both the f110w and f160w filters ( approximately similar to the bessell _ j _ and _ h _ filters , respectively ) in a tiled pattern ; the integrations , of 512s each , were taken in multiaccum mode to preserve dynamic range and to correct for cosmic rays . unfortunately , the separation of real sources from artefacts in m82 b is problematic , due to significant small - scale variations in the amplitude of the background emission , which are largely caused by the highly variable extinction . for this reason , we can not use standard unsharp masking techniques to remove this background , since this produces significant residual emission along the dust features . an initial visual examination of the multi - passband observations revealed a multitude of faint point sources that become increasingly obvious with increasing wavelength . to include in our source selection a maximum number of real and a minimum amount of spurious sources ( due to , e.g. , dust features , weak cosmic rays , or poisson noise in regions of high surface brightness or at the ccd edges ) , we decided to cross correlate source lists obtained in individual passbands . we performed extensive checks to find the best selection criterions and thus to minimize the effects introduced by artefacts on the one hand and the exclusion of either very red or very blue sources on the other . we chose our detection thresholds such that the number of candidate sources selected from the images in all passbands were comparable , of order 4000 . then , we cross correlated the source lists obtained in the individual passbands . finally , we determined which combination of passbands resulted in the optimal matching of sources detected in both the blue and the nir passbands . this source selection procedure lead us to conclude that a final source list obtained from the cross correlation of the candidate sources detected in the f555w and f814w filters would contain the most representative fraction of the m82 b - region source population . however , an initial visual examination of the cross - correlated _ v _ and _ i_-band sources showed that the automated detection routine had returned non - detections as well as artefacts that were clearly associated with dust features . it also showed that the quality of the resulting sources for photometric follow - up was highly variable . therefore , we decided to check the reality of all cross - correlated sources by examining them visually , in both passbands , at the same time classifying them in terms of contrast , sharpness , and the presence of nearby neighbors . moreover , we added ( and verified ) the @xmath4 30% of the sources in each field that were missed by the automated detection routine , but were clearly real sources , and were recognized as such by the visual examination of both the _ v _ and the _ i_-band images . the final source lists thus obtained contain 737 and 642 verified sources in m82 b1 and b2 , respectively . we estimated the completeness of our object list by randomly and uniformly adding 500 synthetic point sources of input magnitudes between 20.0 and 25.0 mag to the observed images in both the f555w and the f814w filters . the psfs of the synthetic point sources were obtained from observational psfs , and scaled to the desired magnitudes . the effects of crowding in our simulated star fields are small : only @xmath4 12% of the simulated objects were not retrieved due to crowding or overlap of adjacent sources . for the uniformly distributed simulated sources , we established the 50% completeness limits at f555w @xmath18 and 23.3 mag ( for b1 and b2 , respectively ) and at f814w @xmath19 mag . in addition to these synthetic star fields , we used our @xmath20cen _ hst _ observations in the f555w passband , obtained as part of program 6053 ( pi oconnell , cycle 5 ) , to verify the reduction procedures and the accuracy of our photometry . we based the distinction between stars and more extended sources in the m82 b fields on the statistical differences between the size characteristics of the populations of real sources in these fields and those of the stars in the @xmath20cen control field . to do so , we added a scaled version of the @xmath20cen field to the m82 b fields , such that the output magnitudes of the majority of the @xmath20cen stars were in the same magnitude range as those of the m82 b sources , in the latter reference frame . we determined characteristic sizes of both our m82 b verified sources and the @xmath20cen stars , using a gaussian fitting routine . although the true luminosity profiles of the star clusters in m82 b may differ from gaussians ( sect . [ profiles.sect ] ) , this method allows us to distinguish between compact and extended sources . a comparison between the distribution of characteristic stellar sizes from the globular cluster field with those of the verified sources in m82 b revealed that the population of extended sources in m82 b is well - represented by sources with @xmath21 pixels . therefore , in the following we will consider those sources with @xmath22 and _ v_-band magnitude brighter than the 50% completeness limit to be part of our verified cluster samples . they contain 128 and 218 cluster candidates in b1 and b2 , respectively . a close examination of the color distributions in both fields reveals that the cluster population in b2 displays a peak at redder colors , whereas the dispersion in colors is larger than in the b1 region , in particular on the redward side . both these observations hint at a more heavily obscured globular cluster population in b2 . under the assumption that the candidate cluster populations in b1 and b2 are coeval , which is likely if they originated in the same starburst , the differences between their color distributions can entirely be attributed to extinction effects , amounting to an excess extinction in b2 of @xmath23 mag . their color - magnitude diagrams ( cmds ) reveal that the brightest young cluster candidates have approximately identical colors and magnitudes in either region : @xmath24 mag , and @xmath25 ; this indicates that these are mainly unaffected by extinction . we take as tentative color limits for the determination of the intrinsic age spread in the m82 b cluster population @xmath26 and @xmath27 . the limits on the blue side are set by the colors of the bluest cluster candidates ; the red limits are the colors of the brightest ( supposedly unobscured ) clusters in either field , since from this color range , most cluster colors can be achieved with up to 2(3 ) magnitudes of visual extinction , which is not an unreasonably large amount , considering the highly variable extinction features seen in the images . employing the bc96 initial burst models yields ages between @xmath28 and @xmath29 yr , assuming solar metallicity . from the optical near - infrared color - color diagrams shown in fig . 1 , we derive a lower limit to the allowed age range for the bulk of the cluster population of 200300 myr , entirely consistent with the age estimates obtained from the optical cmds . in fig . 2 we present the light profiles of the five brightest star cluster candidates in each field . these objects have sufficient s / n ratios and are relatively isolated , thus enabling us to follow the light profiles out to reasonably large radii . several cluster radial luminosity functions have recently been used in the literature , the simplest of them being a two - dimensional gaussian . alternatively , more complex models have been used , that are better representations of the true stellar light distribution in _ globular _ clusters . the most general of these is the mathematically convenient function of surface brightness @xmath30 as a function of radius _ r _ , proposed by elson et al . ( 1987 ) : from these figures , it is obvious that even though all three fitting functions may represent the inner profiles relatively well , the discrepancy between the results in the outer parts is a strong argument in favor of the more extended modified hubble or @xmath34 ( `` elson '' ) profiles . elson et al . ( 1987 ) try to explain the approximate power - law behavior of the lmc cluster profiles at large radii in terms of different star formation time scales . they argue that the observations probably favor slow star formation . under these circumstances _ and _ by assuming a high star formation efficiency , relaxation of a sufficiently clumpy gaseous protocluster through clump - clump or clump - star two - body encounters may produce the observed power - law luminosity profiles . some of the star clusters show evidence for asymmetrical structure , which may mean that they contain subclumps , or have not yet reached dynamical equilibrium . a visual inspection of these clusters shows that they are noticeably elongated . in the case of these clusters in particular , we may be witnessing ongoing merging processes . however , to strengthen this suggestion , we will need spectroscopic follow - up observations . similar subclustering has been observed in the super star clusters in ngc 1569 ( oconnell et al . 1994 ) and some of the rich lmc star clusters ( e.g. , fischer et al . 1993 ) . bruzual , g. , charlot , s. , 1996 , in : leitherer , c. , alloin , d. , fritze - von alvensleben , u. , et al . , 1996 , _ pasp _ * 108 * , 996 carlson , m.n . , holtzman , j.a . , watson , a.m. , et al . , 1998 , _ aj _ * 115 * , 1778 conti , p.s . , leitherer , c. , vacca , w.d . , 1996 , _ apj _ * 461 * , l87 elson , r.a.w . , fall , s.m . , freeman , k.c . , 1987 , _ apj _ * 323 * , 54 fischer , p. , welch , d.l . , mateo , m. , 1993 , _ aj _ * 105 * , 938 ho , l.c . , 1997 , in : starburst activity in galaxies , eds . franco , j. , terlevich , r. , tenorio - tagle , g. , rev . a. a. conf . series , vol . 6 , mexico city : iaunam , p. 5 , filippenko , a.v . , 1996 , _ apj _ 472 , 600 holtzman , j.a . , faber , s.m . , shaya , e.j . , et al . , 1992 , _ aj _ 103 , 691 marcum , p. , oconnell , r.w . , 1996 , in : from stars to galaxies : the impact of stellar physics on galaxy evolution , eds . leitherer , c. , fritze - von alvensleben , u. , huchra , j. , san francisco : asp , p. 419 meurer , g.r . , 1995 , _ nat _ * 375 * , 742 oconnell , r.w . , gallagher , j.s . , hunter , d.a . , 1994 , _ apj _ * 433 * , 65 oconnell , r.w . , gallagher , j.s . , hunter , d.a . , colley , w.n . , 1995 , _ apj _ * 446 * , l1 oconnell , r.w . , mangano , j.j . , 1978 , _ apj _ * 221 * , 62 rieke , g.h . , loken , k. , rieke , m.j . , tamblyn , p. , 1993 , _ apj _ * 412 * , 99 telesco , c.m . , 1988 , _ ara&a _ * 26 * , 343 van den bergh , s. , 1995 , _ nat . _ * 374 * , 215 watson , a.m. , gallagher , j.s . , holtzman , j.a . , et al . , 1996 , _ aj _ * 112 * , 534 whitmore , b.c . , schweizer , f. , leitherer , c. , borne , k. , robert , c. , 1993 , _ aj _ * 106 * , 1354	we present high - resolution optical and near - infrared _ hst _ observations of two adjacent regions in the fossil starburst region in m82 , m82 b1 and b2 . the presence of both the active and the fossil starburst in m82 provides a unique physical environment to study the stellar and dynamical evolution of star cluster systems . + the cluster population in b2 is more heavily affected by internal extinction than that in b1 , amounting to an excess extinction in b2 of @xmath0 mag . preliminary age estimates date the cluster population in the fossil starburst between @xmath1 and @xmath2 years . the radial luminosity profiles of the brightest clusters are more closely approximated by power laws than by a gaussian model , in particular in their wings , which favors a slow star formation scenario .
the galactic zone of avoidance ( zoa ) presents an obstacle to our efforts to understand the dynamics of the ` local ' universe . namely , the milky way prevents a full - sky mapping of the light , and therefore mass , distribution that induces the local velocity field . thus , a prime motivation of the astronomical study of the zoa is to survey the galaxy distribution in the regions obscured by the milky way . these efforts have indeed uncovered a very rich structure in the zoa , in particular in the region of the great attractor ( ga ; kraan - korteweg & lahav 2000 , and references therein ) . in spite of all the efforts and successes in unveiling the zoa , the fact remains that the zoa is a region more difficult to observe than other parts of the sky . it follows that the reconstruction of the large - scale structure ( lss ) in the zoa should rely more on sophisticated statistical methods than other fields of the sky . here , we shall report on a bayesian approach to the reconstruction of the lss from incomplete , sparse and noisy data sets , such as peculiar velocities and redshift surveys . most of the analysis centers on the recovery of the lss from noisy surveys of radial velocities with very incomplete sky coverage , with some analysis of the iras 1.2 jy redshift survey that is presented as well . the main focus of the present paper is on the general methodology with an emphasize on the tools and methods relevant to the reconstruction of the lss . the method is applied to the mark iii ( willick et al . 1995 ) , sfi ( da costa et al . 1996 ) and the enear ( da costa et al . 2000 ) surveys of the radial velocities , as well as the iras 1.2jy redshift survey . most of the work presented here has been done in collaboration mostly with s. zaroubi and the cosmographical implications of this work are presented in zaroubi ( these proceedings ) . here , the general bayesian methodology is presented in 2 . the cosmological parameters estimated from radial velocities surveys are summarized in 3 . the lss and the zoa that is reconstructed from radial velocity catalogs and the iras 1.2jy redshift survey are presented in 4 . a new algorithm of recovering the tidal field from the full velocity field enables the studying of the lss on scales beyond those that are directly probed by the data ( 5 ) . a technique of simulating the dynamical role of the zoa based on n - body simulations that are constrained by the observed lss is described in 6 . a general discussion concludes this review . the general methodology of analyzing surveys of the lss , such as redshift and radial velocity surveys , was extensively presented in zaroubi et al . the main result of the paper is that for gaussian random fields an optimal estimation of the cosmological parameters is provided by a maximum likelihood analysis . the underlying density and velocity fields are reconstructed by a wiener filter ( wf ) applied to the data . monte carlo realizations of the scatter of the residual from the wf fields are done by means of constrained realizations ( hoffman & ribak 1991 ; cr ) . a brief summary is presented here . consider a data set of radial velocities @xmath0 , where @xmath1 @xmath2 is the three dimensional velocity , @xmath3 is the position of the i - th data point and @xmath4 is the statistical error associated with the i - th radial velocity . the assumption made here is of a cosmological model that describes the data well , that systematic errors have been properly dealt with and that the statistical errors are well understood . the data auto - covariance matrix is then written as : @xmath5 ( here @xmath6 denotes an ensemble average . ) the last term is the error covariance matrix . the velocity covariance tensor is calculated within the linear theory . the likelihood of the @xmath7 data points given a model is , @xmath8^{-1/2 } \exp\left ( -{1\over 2}\sum_{i , j}^n { u_i r_{ij}^{-1 } u_j}\right ) . \label{eq : like}\ ] ] an unbiased estimation of the cosmological parameters is given by a maximum likelihood analysis which finds the most probable parameters given the data and within a model space . the general application of the wf / cr method to the reconstruction of lss is described in zaroubi et al . ( 1995 ) , where the theoretical foundation is discussed in relation to other methods of estimation , such as maximum entropy . the specific application of the wf / cr method to peculiar velocity data sets has been presented in zaroubi et al . ( 1999 ) . here we provide only a brief description of the wf / cr method and the interested reader is referred to the last two references for more details . we assume that the peculiar velocity field @xmath9 and the density fluctuation field @xmath10 are related via the linear gravitational - instability theory . under the assumption of a specific theoretical prior for the power spectrum @xmath11 of the underlying density field , one can write the wf minimum - variance estimator of the fields as @xmath12 and @xmath13 a well known problem of the wf is that it attenuates the estimator to zero in regions where the noise dominates . the reconstructed mean field is thus statistically inhomogeneous . in order to recover statistical homogeneity we produce constrained realizations ( cr ) , in which random realizations of the residual from the mean are generated such that they are statistically consistent both with the data and the _ prior _ model ( hoffman & ribak 1991 ) . in regions dominated by good quality data , the crs are dominated by the data , while in the limit of no data the realizations are practically unconstrained . a maximum likelihood analysis has been performed on the mark iii ( zaroubi et al . 1997 ) , sfi ( freudling et al . 1999 ) and enear ( zaroubi et al . 2000 ) surveys of radial velocities . considering the constraints set on hubble s constant ( expressed by @xmath14kms@xmath15mpc@xmath15 ) and the total matter density parameter ( @xmath16 ) , the data does not constrain each parameter seperately but rather a certain non - linear combination of the two . for the cobe - normalized flat @xmath17cdm models , the most probable parameters are given by : @xmath18 @xmath19 @xmath20 all quoted uncertainties refer to @xmath21 confidence levels . for observables that are drawn from a gaussian random field , a goodness - of - fit criterion for a model to be consistent with the data is that the total @xmath22 per degree of freedom should be close to unity . indeed , for all velocity surveys considered , the most likely models obey this criterion . however , given the complex nature of the data ( noise , selection etc . ) the goodness - of - fit criterion has been extended and performed on a mode - by - mode basis ( hoffman & zaroubi 2000 ) . the idea is to use the representation of the data by the eigenmodes of the data auto - covariance matrix . the analysis reveals that for all surveys there is a systematic trend of the @xmath22 to increase with the mode number ( where the modes are sorted by decreasing value of the eigenvalues ) . this suggests a systematic inconsistency of the assumed theoretical model and/or error analysis with the data . this trend is statistically very significant in the case of the mark iii and sfi surveys , and less so in the case of enear . the wf / cr algorithm has been applied both to mark iii ( zaroubi et al . 1999 ) , sfi ( hoffman & zaroubi , unpublished ) and the enear ( zaroubi et al . 2000 ) surveys . the lss uncovered by the velocity surveys is shown in fig . 1 , where the density and velocity evaluated at the supergalactic plane are shown . a view of that structure in galactic aitoff projection , focusing on the zoa , is presented in zaroubi ( these proceedings ) . a comparison of the structure revealed by the three structure shows a gross agreement with respect to the main ` players ' in the local dynamics , however the fine structure differs . inspection of the large - scale velocity field reveals that in all cases the flow is dominated by two main structures , the ga and the perseus - pisces ( pp ) supercluster . however , the mark iii exhibits a long range coherent component of the velocity field which is clearly induced by distant structures , not directly probed by the data . this component is smaller in the sfi case and is missing altogether from the enear velocity field . the wf / cr has been applied to the iras 1.2jy redshift survey ( bistolas & hoffman 1998 , bistolas 1999 ) . a thorough analysis of the lss inferred by this survey is beyond the scope of the present paper . only the zoa reconstruction is presented here . a study of the cosmography unveiled in the zoa is to be reported elsewhere ( bistolas , kraan - korteweg , & hoffman , in preparation ) . here , the zoa is shown in aitoff projections of spherical slices ranging from @xmath23 to @xmath24kms@xmath15 ( figs . 2 and 3 ) . the velocity field can be decomposed into two components , one which is induced by the local mass distribution and a tidal component due the external mass . here , we follow the procedure suggested by hoffman ( 1998a , b ) and more recently by hoffman et al . the key idea is to solve for the particular solution of the poisson equation with respect to the wf density field within a given region and zero padding outside . this yields the velocity field induced locally , hereafter the divergent field . the residual of the full velocity field is the the tidal field . figure 4 shows the decomposition applied to the mark iii survey , where the local volume is a sphere of 6000kms@xmath15 centered on the local group . the plots show the full velocity field , the divergent and tidal components . to further understand the nature of the tidal field , its bulk velocity component has been subtracted and the residual is shown as well . this residual is clearly dominated by a quadrupole component . in principle , the analysis of this residual field can shed light on the exterior mass distribution . indeed , the full velocity field recovers the cmb dipole velocity , half of which is contributed by the tidal field that is induced by structures outside of the 6000kms@xmath15 sphere . the tidal field is decomposed into a dipole and quadrupole moments , namely a bulk velocity and traceless shear tensor . the directions of the bulk velocity and the eigenvectors of the shear tensor are shown in aitoff projection in galactic @xmath25 coordinates ( fig . the figure shows that both the bulk velocity and the dilational eigenmode lie in , or close to , the zoa . the tidal field has a bulk velocity with an amplitude of @xmath26kms@xmath15 in the direction of ( galactic ) @xmath27 . it follows that the deep zoa ( on scales larger than 6000kms@xmath15 ) induces a substantial part of the local dynamics . the wf / cr is formulated within the linear theory of the gravitational instability , where the deviations from a pure friedmann model constitute a gaussian random field . however , as these deviations evolve away from the linear regime , the use of linear tools is of very limited value . this algorithm can be extended to provide quasi non - linear crs . the basic idea is to use constraints that are drawn from data that is linear or quasi - linear , to generate linear crs given a _ prior _ model , take it backwards in time and feed these realizations as initial conditions to n - body simulations . the final configuration is obtained by solving exactly the non - linear dynamics subject to the boundary conditions imposed by the actual universe . this provides realistic n - body simulations that reproduce the observed universe , e.g. the ga , pp and the local void with the right location and amplitude . such simulation are essential for non - linear modeling of the dynamical role of the zoa in surveys of the lss . an application of this was done by bistolas & hoffman ( 1998 ) who used the iras 1.2 redshift survey as the constraining data . the resulting structure reproduces the observed galaxy distribution out to a distance of @xmath28kms@xmath15(fig . this procedure has been applied also to the mark iii data ( van de weygaert & hoffman 2000 ; klypin , van de weygaert , & hoffman , in preparation ) . the recovery of the lss in the zoa poses a challenge to our understanding of the dynamics of our ` local ' universe . direct observational efforts of unveiling the zoa , sophisticated as they might be , need to be complemented by elaborate statistical method for the reconstruction of the lss that is hidden in the zoa . within the standard cosmological model in which structure emerges from a primordial gaussian perturbation field _ via _ gravitational instability , the bayesian approach presented here provides the optimal tools for estimating the cosmological parameters and the reconstruction of the lss . the algorithm presented here consists of the following four steps . first , given a data set , the cosmological ( and other ) parameters are estimated by maximum likelihood analysis over a range of models . then , the consistency of the most probable model with the data is analyzed by the differential @xmath22 analysis . the model that is found by the maximum likelihood and the goodness - of - fit analysis , serves as the _ prior _ model for the wf / cr reconstruction . applying the wf to the data , one gets the mean density and velocity fields , thus providing an estimator of the lss . the crs constitute an ensemble of monte carlo simulations that are designed to reproduce the imposed constraints , namely the observed radial velocity and/or redshift surveys . the scatter of the crs around the wf field provides a measure to the robustness of wf reconstruction . a detailed study of the cosmography recovered from the radial velocity surveys is given by zaroubi ( these proceedings ) . here , this has been extended to study structure beyond the depth of these surveys . this is done by the tidal field decomposition of the velocity field . the tidal field induced by the structure beyond a sphere of radius 6000kms@xmath15 has a dipole component of 300kms@xmath15 towards ( galactic ) @xmath29 . this direction implies that a substantial fraction of the structure that induces this dipole is hidden in the zoa . this provides a further motivation for studying the zoa . the present review is based mostly on work done in collaboration with s. zaroubi . a. eldar is gratefully acknowledged for his help with producing some of the figures . i am grateful to the local organizing committee , and in particular rene kraan - korteweg , for organizing a very exciting meeting . this research has been partially supported by the binational science foundation ( 94 - 00185 ) and the israel science foundation ( 103/98 ) .	the uncovering of the large - scale structure that is hidden in the zone of avoidance is essential to our quest for the understanding of the dynamics of the nearby universe . dedicated and sophisticated observations of the zone of avoidance have to be matched by equally sophisticated theoretical tools for the reconstruction of the large - scale structure from sparse , noisy and very incomplete data . a general approach for the estimation of cosmological parameters and the reconstruction of the density and velocity fields is presented here , based on maximum likelihood analysis , wiener filtering and constrained realizations . radial velocity surveys and the iras 1.2jy redshift survey are used to reconstruct the zone of avoidance . # 1_#1 _ # 1_#1 _ = # 1 1.25 in .125 in .25 in # 1 1.25 in .125 in .25 in
several suggestions are already present about the possibility of interesting isospin effects on the transition to a mixed hadron - quark phase at high baryon density @xcite . this seems to be a very appealing physics program for the new facilities , fair at gsi - darmstadt @xcite and nica at jinr - dubna @xcite , where heavy ion beams ( even unstable , with large isospin asymmetry ) will be available with good intensities in the 1 - 30 agev energy region . the weak point of those predictions is the lack of a reliable equation of state that can describe with the same confidence the two phases , hadronic and deconfined . on the other hand this also represents a strong theory motivation to work on more refined effective theories for a strong interacting matter . the main qualitative argument in favor of noticeable isospin effects on the hadron - quark transition at high density can be derived from the fig.[isoparton ] , where we compare typical equations of state ( eos ) for hadron ( nucleon ) and quark matter , at zero temperature , for symmetric ( @xmath2 ) and neutron matter ( @xmath3 ) , where @xmath4 are the neutron / proton densities and @xmath5 the total baryon density . for the hadron part we use a relativistic mean field ( rmf ) eos ( @xcite ) with non - linear terms and effective @xmath6 coupling for the isovector part , largely used to study isospin effects in relativistic heavy ion collisions @xcite . in order to keep a smooth flow of the physics points in the discussion , details about the adopted effective nucleon - meson lagrangians are presented in the appendix a. of symmetric / neutron matter : hadron ( @xmath7 ) , solid lines , vs. quark ( mit - bag ) , dashed lines . @xmath8 represent the isospin asymmetry parameters respectively of the hadron , quark matter : @xmath9 , symmetric matter ; @xmath10 , neutron matter . ] the energy density and the pressure for the quark phase are given by the mit bag model @xcite ( two - flavor case ) and read , respectively : @xmath11 @xmath12 where b denotes the bag constant ( the bag pressure ) , taken as a rather standard value from the hadron spectra ( @xmath13 , no density dependence ) , @xmath14 are the quark masses ( @xmath15 choice ) , and @xmath16 , @xmath17 represent the fermi distribution functions for quarks and anti - quarks . the quark number density is given by @xmath18 the transition to the more repulsive quark matter will appear around the crossing points of the two eos . we see that such crossing for symmetric matter ( @xmath19 ) is located at rather high density , @xmath20 , while for pure neutron matter ( @xmath21 ) it is moving down to about three times @xmath22 . of course the fig.1 represents just a simple energetic argument to support the hadron - quark transition to occur at lower baryon densities for more isospin asymmetric matter . in the rest of the paper we will rigourously consider the case of a first order phase transition in the gibbs frame for a system with two conserved charges ( baryon and isospin ) , in order to derive more detailed results . since the first order phase transition presents a jump in the energy , we can expect the mixed phase to start at densities even before the crossing points of the fig.1 . the lower boundary then can be predicted at relatively low baryon densities for asymmetric matter , likely reached in relativistic heavy ion collisions . moreover this point is certainly of interest for the structure of the crust and the inner core of neutron stars , e.g. see refs . @xcite and the review @xcite . the ref.@xcite is particularly interesting since similar results are obtained with rather different hadronic approaches , the rmf and the non - relativistic brueckner - hartree - fock ( bhf ) theory . we finally note that the above conclusions are rather independent on the isoscalar part of the used hadron eos at high density , that is chosen to be rather soft in agreement with collective flow and kaon production data @xcite . in the used bag model no gluon interactions , the @xmath23-strong coupling parameter , are included . we remark that this in fact would enhance the above effect , since it represents an attractive correction for a fixed b - constant , see @xcite . a reduction of the bag - constant with increasing baryon density , as suggested by various models , see ref.@xcite , will also go in the direction of an `` earlier '' ( lower density ) transition , as already seen in ref.@xcite . at variance , the presence of explicit isovector contributions in the quark phase could play an important role , as shown in the following also for other isospin properties inside the mixed phase . we can study in detail the isospin dependence of the transition densities @xcite . the structure of the mixed phase is obtained by imposing the gibbs conditions @xcite for chemical potentials and pressure and by requiring the conservation of the total baryon and isospin densities : @xmath24 where @xmath25 is the fraction of quark matter in the mixed phase and t is the temperature . the consistent definitions for the densities and chemical potentials in the two phases are given by : @xmath26 for the hadron phase and @xmath27 for the quark phase . the related asymmetry parameters are : @xmath28 nucleon and quark chemical potentials , as well as the pressures in the two phases , are directly derived from the respective eos . in this way we get the @xmath29 surface which gives the phase coexistence region in the @xmath30 space . for a fixed value of the total asymmetry @xmath31 we will study the boundaries of the mixed phase region in the @xmath32 plane . since in general the charge chemical potential is related to the symmetry term of the eos , @xcite , @xmath33 we expect critical and transition densities rather sensitive to the isovector channel in the two phases . in the hadron sector we will use the non - linear relativistic mean field models , @xcite , with different structure of the isovector part , already tested to describe the isospin dependence of collective flows and meson production for heavy ion collisions at intermediate energies , @xcite . we will refer to these different iso - lagrangians as : i ) @xmath34 , where no isovector meson is included and the symmetry term is only given by the kinetic fermi contribution , ii ) @xmath35 when the interaction contribution of an isovector - vector meson is considered and finally iii ) @xmath36 where also the contribution of an isovector - scalar meson is accounted for . see details in appendix a and refs.@xcite . we will look at the effect on the hadron - quark transition of the different stiffness of the symmetry term at high baryon densities in the different parametrizations . as clearly shown in appendix a , where a rather transparent form for the density dependence of the symmetry energy in rmf approaches is discussed , the potential part of the symmetry term will be proportional to the baryon density in the @xmath35 choice and even stiffer in the @xmath36 case . as already mentioned , in the quark phase we use the mit - bag model , where the symmetry term is only given by the fermi contribution . the bag parameter b is fixed for each baryon density to a constant , rather standard , value @xmath37 , corresponding to a bag pressure of @xmath38 . ) and upper ( @xmath39 ) boundaries of the mixed phase , at zero temperature , vs. the asymmetry parameter . quark @xmath40 : @xmath41 bag model with @xmath42=160 @xmath43 . ] 1.0 cm [ boundaries ] in general for each effective interactive lagrangian we can simulate the solution of the highly non - linear system of eqs.([gibbs ] ) , via an iterative minimization procedure , in order to determine the binodal boundaries and the critical end point ( cep ) ( @xmath44 ) of the mixed phase . a relatively simple calculation can be performed at zero temperature . the isospin effect ( asymmetry dependence ) on the lower @xmath45 and upper @xmath46 transition densities of the mixed phase are shown in fig.[boundaries ] for various choices of the hadron eos . the effect of a larger repulsion of the symmetry energy in the hadron sector , from @xmath34 to @xmath35 and to @xmath36 , is clearly evident on the lower boundary with a sharp decrease of the transition density even at relatively low asymmetries . ) and asymmetric ( @xmath47 ) matter . hadron eos from @xmath35 interaction . quark @xmath40 : @xmath41 bag model with @xmath42=160 @xmath43 . the grey region corresponds to the conditions reached in heavy ion collisions simulations at few agev beam energies , see section 3 . ] typical results for isospin effects on the whole binodal `` surface '' are presented in fig.[tcrit ] for symmetric and asymmetric matter . for the hadron part we have started from a @xmath35 effective lagrangian very close to other widely used relativistic effective models , e.g. see the @xmath48 of ref.@xcite and the @xmath49 interaction of p.ring and collaborators @xcite , which has also given good nuclear structure results , even for exotic nuclei . in the symmetric matter case the mixed phase is evaluated from the simpler maxwell conditions . the results are shown in fig.[maxwell ] for the same hadron and quark @xmath40 s as in fig.[tcrit ] at temperatures t=0 , 50 and 80 mev . the equal chemical potential densities ( intersection of the dotted line in the lower panel ) must correspond to the equal pressure densities of the upper panels . we nicely see : i ) at t=0 mev the mixed phase is centered around @xmath50 , exactly the @xmath51 crossing point of fig.[isoparton ] , confirming our energetic argument about the transition location ; ii ) the size of the mixed phase is shrinking with temperature , it is very narrow at t=50 mev and finally at t=80 mev we can not have anymore a first order transition . in fact the critical end point is found at @xmath52 , @xmath53 , @xmath54 and @xmath55 , see also fig.[tcrit ] . the result is dependent on the choice of the bag constant , with increasing of the critical temperature with the bag value due to the reduction of the pressure in the quark phase , while the chemical potentials are not affected . this point will not be further discussed since here the main focus is on the isospin dependence of the mixed phase at lower temperatures , that can be probed in heavy ion collisions at intermediate energies . as expected , the lower boundary of the mixed phase is mostly affected by isospin effects . in spite of the relatively small total asymmetry , @xmath47 , we clearly observe in fig.[tcrit ] a shift to the left of the first transition boundary , in particular at low temperature , and an indication of a relatively `` earlier '' critical end point , around ( @xmath56 ) . actually we see from fig.[tcrit ] that for asymmetric matter , we are not able to explicitly reach the @xmath57 of the binodal surface . since this feature appears also in the following figures we must add a general comment . from the present results we can not exclude a @xmath57 at higher temperature and smaller baryon density . however we have also to mention some numerical problems . the solution of the gibbs conditions , the highly non - linear system of eqs.([gibbs ] ) , is found through an iterative multi - parameter minimization procedure ( the newton - raphson method ) . when we are close to the critical end point , for each @xmath25-concentration the baryon and isospin densities of the hadron and quark phases become very similar and we start to have problems in finding a definite minimum . we are working on improving the numerical accuracy but in any case the trends towards the @xmath57 are always quite clearly observed , as a severe reduction of the mixed phase . in the following we will concentrate on properties of the mixed phase mostly located at high density and relatively low temperature , well described in the calculation and within the reach of heavy ion collisions in the few agev range , see section 3 . for @xmath47 asymmetric matter , in the figs.[nlrmix ] , [ nlrdmix ] we show also the ( @xmath0 ) curves inside the mixed phase corresponding to a @xmath58 and @xmath59 presence of the quark component ( @xmath60 ) , evaluated respectively with the two choices , @xmath35 and @xmath36 , of the symmetry interaction in the hadron sector . we note , as also expected from fig.[boundaries ] , that in the more repulsive @xmath36 case the lower boundary is much shifted to the left . however this effect is not so evident for the curve corresponding to a @xmath58 quark concentration , and almost absent for the @xmath59 case . the conclusion seems to be that for a stiffer symmetry term in a heavy - ion collision at intermediate energies during the compression stage we can have more chance to probe the mixed phase , although in a region with small weight of the quark component . binodal surface and ( @xmath0 ) curves for various quark concentrations ( @xmath60 ) in the mixed phase . quark @xmath40 : @xmath41 bag model with @xmath42=160 @xmath43 . hadron @xmath40 : @xmath35 effective interaction . ] , for the @xmath36 effective interaction in the hadron sector . ] binodal surface and ( @xmath61 ) curves for various quark concentrations in the mixed phase . quark @xmath40 : @xmath41 bag model with @xmath42=160 @xmath43 . hadron @xmath40 : @xmath35 effective interaction . ] , for the @xmath36 effective interaction in the hadron sector . ] in fact from the solution of the system eq.([gibbs ] ) we get the baryon densities @xmath62 in the two phases for any @xmath25 value . in the figs . [ nlrmixhq ] , [ nlrdmixhq ] we present the results for the same weights @xmath63 of the quark phase of the previous figures . the quark phase appears always with larger baryon density , even for the lowest value of the concentration . can we expect some signatures related to the subsequent hadronization in the following expansion ? and @xmath47 . @xmath35 and @xmath36 effective hadron interactions are considered . quark @xmath40 : @xmath41 bag model with @xmath42=160 @xmath43 . ] an interesting possibility is coming from the study of the asymmetry @xmath64 in the quark phase . in fact since the symmetry energy is rather different in the two phases we can expect an isospin distillation ( or fractionation ) , very similar to the one observed in the liquid - gas transition in dilute nuclear matter @xcite , this time with the larger isospin content in the higher density quark phase . in fig.[alphaq ] we show the asymmetry @xmath64 in the quark phase as a function of the quark concentration @xmath25 for the case with global asymmetry @xmath47 ( zero temperature ) . the calculation is performed with the two choices of the symmetry term in the hadron sector . we see an impressive increase of the quark asymmetry when we approach the lower boundary of the mixed phase , even to values larger than one , likely just for numerical accuracy @xcite . of course the quark asymmetry recovers the global value 0.2 at the upper boundary @xmath65 . a simple algebraic calculation allows to evaluate the corresponding asymmetries of the hadron phase . in fact from the charge conservation we have that for any @xmath25-mixture the global asymmetry @xmath66 is given by : @xmath67 for any @xmath25 , from the calculated @xmath64 of fig.[alphaq ] and the @xmath62 of figs.[nlrmixhq ] , [ nlrdmixhq ] , we can get the correspondent asymmetry of the hadron phase @xmath68 . for a @xmath58 quark concentration we have an @xmath69 ratio around 5 for @xmath35 and around 20 for @xmath36 , more repulsive in the isovector channel . it is also interesting to compare the isospin content @xmath70 of the high density region expected from transport simulations without the hadron - quark transition and the effective @xmath70 of the quark phase in a @xmath58 concentration . in the case of @xmath71 ( initial @xmath72 ) central collisions at @xmath73 in pure hadronic simulations we get in the high density phase a reduced @xmath74 ( respectively with @xmath75 interactions ) due to the fast neutron emission @xcite . the corresponding isospin content of the quark phases is much larger , @xmath76 for @xmath35 and @xmath77 for @xmath36 . this is the @xmath78 effect discussed in the section 3 . we could expect a signal of such large asymmetries , coupled to a larger baryon density in the quark phase , in the subsequent hadronization . we finally remark that at higher temperature and smaller baryon chemical potential ( ultrarelativistic collisions ) the isospin effects discussed here are expected to vanish @xcite , even if other physics can enter the game and charge asymmetry effects are predicted also at @xmath79 and @xmath80 @xcite . all the above results will be very sensitive to the explicit inclusion of isovector interactions in effective non - perturbative qcd models in regions at high baryon chemical potentials . unfortunately few attempts have been worked out for two main reasons : i ) the difficulties of lattice - qcd calculations at high baryon densities ; ii ) the main interest on the qgp phase transition at high temperature and small baryon chemical potentials , as probed in the expanding fireball of ultrarelativistic heavy ion collisions . a first approach can be supplied by a two - flavor nambu - jona lasinio ( @xmath81 ) model @xcite , which in fact describes the chiral restoration but not the deconfinement dynamics . the isospin asymmetry can be included in a flavor - mixing picture @xcite , corresponding to different couplings to the ( u , d ) quark - antiquark condensates . as a consequence we can have now a dependence of the constituent mass of a given flavor to both quark condensates . we devote the appendix b to a detailed study of this isospin effects in the @xmath81 chiral dynamics . due to the scalar nature of the interacting part of the corresponding lagrangians only the quark effective mass dynamics will be affected . in the `` realistic '' small mixing case , see also @xcite , we get a definite @xmath82 splitting at high baryon density ( before the chiral restoration ) . all that can indicate a more fundamental confirmation of the @xmath83 splitting in the hadron phase , as suggested by the effective @xmath84 model with the isovector scalar @xmath85 coupling , see @xcite . however such isospin mixing effect results in a very small variation of the symmetry energy in the quark phase , related only to the fermi kinetic contribution . moreover we remind that confinement is still missing in this @xmath81 mean field approach . in any case there are extensive suggestions about a favored chiral symmetry restoration in systems with large neutron excess @xcite . more generally starting from the qcd lagrangian one can arrive to an effective color current - current interaction where an expansion in various components can provide isovector contributions , @xcite . in this respect we remark another interesting `` indirect '' isospin effect , i.e. not directly related to isovector terms in the effective lagrangian , related to the presence of quark condensates due to the attractive gluon interaction . we note that just a color - pairing effect in the two - flavor system ( the @xmath86 phase @xcite ) would imply a stiffer symmetry energy in the quark eos since we have a larger attraction when the densities of up and down quarks are equal . a first study of the high density hadron - quark transition including such gluon correlation in the bag model has been presented very recently in @xcite . now the symmetry energy difference between hadron and quark phases is partially reduced , at least at low temperatures , and consequently also the isospin effects discussed in detail in this work will be less robust , although still present . an interesting point is that in any case the quark phase is more bound due to the attractive gluon contribution . hence the transition to the mixed phase will still appear at relatively low baryon densities , now for an `` isoscalar '' mechanism , within the reach of `` low energy '' heavy ion collisions , i.e. in the range of few agev . as an intuitive picture we can refer again to the fig.1 . essentially the difference between the @xmath87 and @xmath88 curves is increasing but meanwhile both are decreasing . with increasing temperature the color pairing effect will be in general reduced , as confirmed in @xcite in an extended @xmath81 calculation , and so isospin effects , as discussed before , will be more relevant . all that , as well as the location of the critical - end - point , is naturally related to the used value of the superconducting gap , opening new stimulating perspectives . in this sense new experiments on mixed phase properties observed with isospin asymmetric heavy ion collisions , as suggested in the final section , will be extremely important . based on the qualitative argument of the introduction and on more detailed calculations in a first order phase transition scheme , we have predicted rather `` robust '' isospin effects on the hadron - quark transition at high baryon densities , not depending on details of the eos parametrizations in the hadron and quark phases . our results seem to indicate a specific region where the onset of the mixed phase should be mainly located : @xmath89 , @xmath90 , for realistic asymmetries @xmath91 . a key question is if such a region of the phase space can be explored by means of heavy - ion - collisions . in refs.@xcite it is shown that even collisions of stable nuclei at intermediate energies ( @xmath92 ) make available the pertinent ( @xmath93 ) region where the phase transition is expected to occur . in this respect we can refer to the reaction @xmath94 ( average isospin asymmetry @xmath95 ) at @xmath96 that has been investigated in ref.@xcite , using a consistent relativistic mean field approach with the same interactions , for a semicentral impact parameter @xmath97 , chosen just to increase the neutron excess in the interacting region . the evolution of momentum distribution and baryon / isospin densities in a space cell located in the c.m . of the system has been also studied . after about @xmath98 a local equilibration is achieved still in the compressed phase , before the fast expansion . we have a unique fermi distribution and from a simple fit the `` local '' temperature can be evaluated . a rather exotic nuclear matter is formed in a transient time of the order of @xmath99 , with baryon density around @xmath100 , temperature @xmath101 , and isospin asymmetry between @xmath102 and @xmath103 , likely inside the estimated mixed phase region . we note that high intensity @xmath104 beams in this energy range will be available in the first stage of the fair facility @xcite and also at jinr - dubna in the nuclotron first step of the nica project @xcite . which are the observable effects to look at if we enter and/or cross the mixed phase ? as already stressed , a first expectation will be the isospin distillation effect , a kind of _ neutron trapping _ in the quark phase , supported by statistical fluctuations @xcite as well as by a symmetry energy difference in the two phases , as discussed in section 2.1 . in fact while in the pure hadron matter ( neutron - rich ) at high density we have a large neutron potential repulsion ( in both @xmath35 , @xmath36 cases ) , in the quark phase the @xmath105-quarks see a much smaller symmetry repulsion only due to the kinetic contribution from the fermi gas . as a consequence while in a pure hadronic phase neutrons are quickly emitted or `` transformed '' in protons by inelastic collisions @xcite , when the mixed phase starts forming , neutrons are kept in the interacting system , in the quark phase which is also at larger baryon density , up to the subsequent hadronization in the expansion stage @xcite . observables related to such neutron `` trapping '' could be an inversion in the trend of the formation of neutron rich clusters , an enhancement of the production of isospin - rich nucleon resonances and subsequent decays , i.e. an increase of @xmath106 , @xmath107 yield ratios for reaction products coming from high density regions , that could be selected looking at large transverse momenta , corresponding to a large radial flow . if such kinetic selection of particles from the mixed phase can really be successful also other potential signatures would become available . one is related to the general softening of the matter , due to the contribution of more degrees of freedom , that should show up in the damping of collective flows @xcite . the azimuthal distributions ( elliptic flows ) will be particularly affected since particles mostly retain their high transverse momenta escaping along directions orthogonal to the reaction plane without suffering much rescattering processes . thus a further signature could be the observation , for the selected particles , of the onset of a quark - number scaling of the elliptic flow : a property of hadronization by quark coalescence that has been predicted and observed at rhic energies , i.e. for the transition at @xmath79 @xcite . we note that all the above results , on the binodal boundaries of the mixed phase and on the isospin distillation are sensitive to the symmetry term in the hadron sector , although the main isospin effects are present for all the parametrizations of the isovector interaction . at variance , for the quark sector the lack of explicit isovector terms could strongly affect the location of the phase transition in asymmetric matter and the related expected observables . in conclusion the aim of this work is twofold : * to stimulate new experiments on isospin effects in heavy ion collisions at intermediate energies ( in a few @xmath108 range ) with attention to the isospin content of produced particles and to elliptic flow properties , in particular for high-@xmath109 selections . * to stimulate more refined models of effective lagrangians for non - perturbative qcd , where isovector channels are consistently accounted for and/or gluon correlations , leading to diquark condensates , can induce symmetry energy effects . this project is supported by the national natural science foundation of china under grant no . 10875160 and the infn of italy . v.b . is grateful for the warm hospitality at laboratori nazionali del sud , infn . this collaboration is supported in part by the romanian ministry for education and research under the cncsis contract pnii id-946/2007 . a lagrangian density with `` minimal '' meson channels and non - linear terms is used . the nuclear interaction is mediated by two isoscalar , the scalar @xmath110 and the vector @xmath111 , and two isovector , the scalar @xmath85 and the vector @xmath112 , mesons . non linear terms are considered only for the @xmath110 contribution to account for the correct compressibility around saturation . constant nucleon - meson couplings are used , chosen to reproduce the saturation properties and to represent a reasonable average of the density dependence predicted by relativistic brueckner - hartree - fock calculations @xcite , see details in refs.@xcite . @xmath113\psi \nonumber \\ & & + \frac{1}{2}(\partial_{\mu}\sigma\partial^{\mu}\sigma - m_{\sigma}^2\sigma^2 ) -u(\phi)+\frac{1}{2}m^2_{\omega}\omega_{\mu } \omega^{\mu } + \frac{1}{2}m^2_{\rho}\vec{b}_{\mu}\cdot\vec{b}^{\mu } \nonumber \\ & & + \frac{1}{2}(\partial_{\mu}\vec{\delta}\cdot\partial^{\mu}\vec{\delta } -m_{\delta}^2\vec{\delta^2 } ) -\frac{1}{4}f_{\mu\nu}f^{\mu\nu } -\frac{1}{4}\vec{g}_{\mu\nu}\vec{g}^{\mu\nu},\end{aligned}\ ] ] where @xmath114 , @xmath115 , and the @xmath116 is the nonlinear potential of @xmath110 meson : @xmath117 . the eos for nuclear matter at finite temperature in the mean - field approximation ( mfa ) is given by @xmath118 @xmath119 where @xmath120 . the nucleon effective masses are defined as @xmath121 the field equations in the relativistic mean field ( rmf ) approach are @xmath122 with the baryon density @xmath123 and @xmath124 , @xmath125 and @xmath126 are the scalar densities for proton and neutron , respectively . the @xmath127 and @xmath128 in eqs.([edens ] , ( [ press ] ) are the fermion and antifermion distribution functions for protons and neutrons ( @xmath129 ) : @xmath130 where the effective chemical potential @xmath131 is determined by the nucleon density @xmath132 @xmath133 while the scalar density @xmath134 , which gives the coupling to the scalar fields is given by @xmath135 note the @xmath136 quenching factor at high baryon density . the @xmath137 are related to the chemical potentials @xmath138 in terms of the vector meson mean fields by the equation @xmath139 the baryon and isospin chemical potentials in the hadron phase can be expressed in terms of the ( p , n ) ones as @xmath140 in presence of the coupling to the two isovector @xmath141-meson fields , the expression for the symmetry energy has a simple transparent form , see @xcite : @xmath142\rho~,\ ] ] where @xmath143 and @xmath144 . now we easily see that in the @xmath36 choice we have a large increase of the symmetry energy at high baryon densities . the potential simmetry term is given by the combination @xmath145 $ ] of the repulsive vector @xmath112 and attractive scalar @xmath85 isovector couplings . thus , when the @xmath85 is included we have to increase the @xmath112-meso coupling in order to reproduce the same asymmetry parameter @xmath146 at saturation . the net effect will be a stiffer symmetry energy at higher baryon densities due to the @xmath147 quenching of the attractive part . 0.7 cm the coupling constants are fixed from good saturation properties and from dirac - brueckner estimations , see the detailed discussions in refs.@xcite . the isoscalar part of the eos is chosen to be rather soft at high densities , see @xcite , in order to satisfy the experimental constraints from collective flows and kaon production in intermediate energy heavy ion collisions @xcite . the coupling constants , @xmath148 , @xmath149 , and the two parameters of the @xmath110 self - interacting terms : @xmath150 and @xmath151 are reported in table 1 . the @xmath152 mass is fixed at @xmath153 . the corresponding properties of nuclear matter are listed in table 2 . here the binding energy is defined @xmath154 . * table 1 . * parameter set . [ cols="^,^,^",options="header " , ] we finally note that these lagrangians have been already used for flow @xcite , pion production @xcite , isospin tracer @xcite and kaon production @xcite calculations for relativistic heavy ion collisions with an overall good agreement to data . from the above discussion it appears ( extremely ) important to include the isospin degree of freedom in any effective qcd dynamics . a first approach can be supplied by a two - flavor nambu - jona lasinio model where the isospin asymmetry can be included in a flavor - mixing picture @xcite . the lagrangian is given by @xmath155 with @xmath156 the free part @xmath157 and the two different interaction part given by @xmath158 % + { \rm det}_f % [ \bar{\psi } ( 1-\gamma_5 ) \psi ] \right\rbrace= \nonumber \\ g_2 \big\ { ( { \bar \psi } \psi)^2 - ( { \bar \psi}\,\vec\tau \psi)^2 - ( { \bar \psi}\,i\gamma_5 \psi)^2 + ( { \bar \psi}\,i\gamma_5\vec\tau \psi)^2 \big\}.\end{aligned}\ ] ] in the mean field approximation the new gap equations are @xmath159 , @xmath160 , where the @xmath161 are the two ( negative ) condensates which are given by @xmath162 and @xmath163 the ( equal ) current masses . introducing explicitily a flavor mixing , i.e. the dependence of the constituent mass of a given flavor to both condensate , via @xmath164 we have the coupled equations @xmath165 for @xmath166 we have back the usual njl ( @xmath167 ) , while small / large mixing is for @xmath168/@xmath169 respectively . the value of @xmath170 has a consequence on the structure of the phase diagram in the region of low temperatures and high chemical potential . in fact as shown in @xcite for @xmath171 there are two distinct phase transitions for the up quarks and for the down quarks , but for this value the interaction is symmetric under @xmath172 transformations and it is unrealistic . while for @xmath173 the @xmath172 symmetry becomes explicitly broken and there is only a single first order phase transition . realistic estimations of @xmath170 fitting the physical @xmath174-meson mass give a value of @xmath175 @xcite . in neutron rich matter @xmath176 decreases more rapidly due to the larger @xmath177 and so @xmath178 . in the `` realistic '' small mixing case we will get a definite @xmath179 splitting at high baryon density ( before the chiral restoration ) . this expectation is confirmed by a full calculation of the coupled gap equations with standard parameters @xcite . all that can indicate a more fundamental confirmation of the @xmath83 splitting in the hadron phase , as suggested by the effective @xmath84 model with the isovector scalar @xmath85 coupling , see @xcite . however such isospin mixing effect results in a very small variation of the symmetry energy in the quark phase , still related only to the fermi kinetic contribution . in fact this represents just a very first step towards a more complete treatment of isovector contributions in effective quark models , of large interest for the discussion of the phase transition at high densities . in principle there is nothing wrong with @xmath182 evaluations ( although we note that for pure `` neutron '' matter the quark phase asymmetry should be 1 , since @xmath183 ) . indeed for low @xmath25-values , @xmath184 , very small quark concentrations , we can get @xmath64 values slightly larger than 1 . however we are cautious about these results since we can expect also some numerical problems . in fact for very small @xmath25 values the weight of the @xmath64 contribution in the minimization procedure is expected not too relevant , as we can clearly see from the eq.([alphamix ] ) . in any case the important point is that this is not affecting the general discussion about the isospin distillation .	the phase transition of hadronic to quark matter at high baryon and isospin density is analyzed . nonlinear relativistic mean field models are used to describe hadronic matter , and the mit bag model is adopted for quark matter . the boundaries of the mixed phase and the related critical points for symmetric and asymmetric matter are obtained . isospin effects appear to be rather significant . the binodal transition line of the ( @xmath0 ) diagram is lowered to a region accessible through heavy ion collisions in the energy range of the new planned facilities , e.g. the @xmath1 projects . some observable effects are suggested , in particular an _ isospin distillation _ mechanism . the dependence of the results on a suitable treatment of isospin contributions in effective qcd lagrangian approaches , at the level of explicit isovector parts and/or quark condensates , is critically discussed . nuclear matter at high baryon density ; symmetry energy ; deconfinement transition ; critical end point ; effective qcd lagrangians + pacs numbers : 05.70.fh , 05.70.ce , 21.65.+f , 12.38mh
photon is one of the oldest elementary particle with very well known ( qed ) properties , and as such it is considered to be a tool to test the structure of more complicated objects like a proton . for example , in the dis@xmath0 , considered as the main source of information on the inner structure of the proton , the structure functions @xmath1 for the proton can be measured . the probe - in a wide range of @xmath2 - is provided by flux of virtual photons , emitted by primary electrons ( positrons ) . -11.9cm-0.7 cm -0.9 cm so , besides proton also electron and photon are involved in the measurement of the structure functions and the parton densities of the proton . a _ factorization _ allows to treat independently the electron and proton part ( figure above ) . the contributions of the longitudinal ( l ) and transverse ( t ) virtual photons can also be _ separated_. therefore , we can write @xmath3 note , that the structure functions @xmath1 are proportional to the combinations of the cross sections for processes @xmath4 . one can ask if the photon is only a in such measurements and to what extend its apparent `` structure '' influences the corresponding structure functions of the proton ? effects of the `` structure '' of the photon may influence our knowledge of the partonic content of the proton also in the large-@xmath5 photoproduction at hera . however here both the initial ( almost real ) photon and the initial proton are on equal footing and it is natural nowdays to take into account partonic content of both of them . similarly , the dis events at hera , if correspond to the large-@xmath5 single particle or jet production for @xmath6 , are sensitive to the `` structure '' of the virtual photon ( with virtuality @xmath2 ) . it is clear , that if we want to understand a proton , we should understand a photon itself . it is interesting to realized that approximately at the same time ( 1922 - 3 ) we have learned that the proton is not a fundamental particle , while the photon is such a particle ( compton effect ) . since early sixties xx we realized that the photon in fact belongs to a `` hadron family '' with the closest relatives ( according to the vmd idea)- the vector mesons:@xmath7 and @xmath8 next , the notion of a partonic content of @xmath9 have been introduced to describe in an effective way its hadronic interaction . at present colliders testing the partonic content of the photon is a difficult task , as at both lep and hera the photon is not a primary particle . the question if it is simpler to deal with a partonic content of the electron was addressed during this workshop . in the inclusive process @xmath10 one can probe the `` structure '' of photon in various ways depending on the virtualities of emitted photons , @xmath11 and @xmath12 ( figure below ) : + @xmath13 the region @xmath14 corresponds to dis@xmath15 , where one can measure the structure functions of the virtual photon . + @xmath13 for @xmath16 , @xmath17 the corresponding process is dis@xmath18 , where one can measure the structure functions of the real photon . + @xmath13 for @xmath19 the total cross section @xmath20 can be studied . -1.2cm,1.2 cm -7.0 cm in the dis events @xmath21 various structure functions for the real photons can be measured in principle , however data are mainly for @xmath22 . the @xmath22 structure function can be related to the partonic densities of the real photon . in the parton model ( pm ) one obtains a prediction from a pure qed that the quark densities @xmath23 , and that @xmath24 are large at the large value of @xmath25 . in qcd there are leading @xmath26 corrections which can be sum up to all orders using an inhomogeneous dglap evolution equations . this form of equation allows to obtain an asymptotic solutions for quark densities in the photon without any input . in practice , one use an input usually based on vmd ( sometimes modelled by @xmath27 ( sic ! ) ) . in many analyses the quark densities in @xmath9 are treated as being proportional to @xmath28 , while in others to @xmath29 , since the main @xmath26-dependence mentioned above has nothing to do with qcd and @xmath30 . the way of counting has practical consequences as it influences the choice of diagrams which define the nlo set . the cross section for the large-@xmath5 hadronic processes involving photons get contributions from diagrams with the direct and the resolved photon interaction with quarks . note , that the resolved - photon contribution are possible for real and virtual photons , in the latter case provided that @xmath31 is larger that @xmath32 - a `` virtuality '' of the @xmath33 . note that for the resolved virtual photons processes one should take into account not only cross section for a defined polarization of @xmath33 ( l or t ) but also , since @xmath9 is not a primary particle , the corresponding interference terms : lt and t1 t2 . there many new data for the proton from hera and for photon from the lep ( dis and large-@xmath5 events ) and from hera ( large-@xmath5 events ) , as presented at this workshop . there is an impressive theoretical progress in the construction of the parton parametrizations for the proton with a higher precision and using various evolution equations , dglap , bfkl and ccfm . of a great importance are special treatments of the heavy - quark thresholds and error analyses of parton densities preformed recently . this stimulates the corresponding progress for a photon , reported during workshop . * in 1934 l. de broglie considered `` the neutrino theory of light '' , with the light quanta composed of @xmath34 pairs ( also pauli , heisenberg , jordan , kronig , born ) . no satisfactory derivation of bose statistics of light quanta from fermi statistics for @xmath35 was found and this idea was abandoned ( almost ) . * the new idea is a non - commutative qed with a selfinteraction of photons . _ the study of light has resulted in achievements of insight , imagination and ingenuity unsurpassed in any field of mental activity ; it illustrates , too , better than any other branch of physics , the vicissitudes of theories _ ( sir j.j . thomson , 1925 ) . for future study a photon collider with a real photon as a primary particle will be very useful . 0.3 cm * acknowledgment : * i would like to thank the local organizing committee for this excellent workshop and all forms of support during my stay in frascati .	a short overview of basics aspects of hadronic interaction of the photon is presented .
noise induced by the environment has been for decades considered the major enemy of quantum technologies . it is nowadays recognised that this initial belief was wrong @xcite . not only can noise be used to generate quantum properties such as entanglement @xcite , but also the dynamics of an open system , e.g. , its coherence time , can be modified by reservoir engineering . generally , there are a number of ways to change the properties of the environment in a selective and controllable way . typical examples are modifications of the spectral density of the electromagnetic field acting as an environment , such as in cavity quantum electrodynamics or photonics band gap materials @xcite , and dynamical decoupling methods @xcite . these techniques are nowadays routinely performed in the laboratories @xcite . at the same time , the experimental ability to isolate from the environment and coherently control individual qubits in solid state systems such as nv centres in diamonds @xcite and superconducting josephson junctions @xcite has made them ideal candidates for quantum technologies . however , despite the advances of the last decade , the effects of noise in these systems still needs to be taken into account to study their robustness , efficiency , and life - time in realistic physical conditions . the rising importance of both reservoir engineering techniques and solid state qubits highlights the need to investigate open quantum systems models which go beyond the markovian approximation usually used in quantum optics . during the last few years research on non - markovian dynamics has flourished . the study of memory effects , characterising non - markovian systems , has been linked with a partial revival of information on the open system @xcite . several ways to quantify information flow and back - flow have been proposed in order to understand the physical phenomena underlying non - markovian evolution @xcite . finally , intense research activity is currently focussed on the understanding of the conditions and the potential advantages of memory effects to enhance the performance of quantum devices @xcite . the main difficulty when dealing with non - markovian models is the lack of a general theorem which guarantees the physicality of the state as time evolves . from a mathematical point of view , one of the reasons why markovian master equations have been so popular is indeed the existence of the gorini - kossakowski - sudarshan - lindblad ( gksl ) theorem characterising completely positive and trace preserving ( cptp ) dynamical maps @xcite . this in turn guarantees that , in absence of initial system - environment correlations , the time evolution of any quantum state of the open system , as described by the solution of the gksl master equation , is always physical . because of this difficulty , dealing with generalised non - markovian master equations is always a tricky business @xcite . even for a single qubit , where conditions for complete positivity are known @xcite , all studies of non - markovianity have mostly focused on very simple models for which an exact solution of the total , i.e. , system plus environment , dynamics is available @xcite . this indeed guarantees physicality by construction . typical examples are the purely dephasing model @xcite , the pauli channel model @xcite and the amplitude damping model @xcite . the latter one goes beyond unital dynamics but is restricted to the case in which the two - state system dissipatively interacts with a zero - temperature reservoir . in this paper we go beyond the existing literature in several ways . firstly , we solve and study the cptp conditions of a generic time - local master equation which contains heating , dissipation and pure dephasing terms . this allows to assess the question of additivity of noise under non - markovian dynamics , extending the results of ref . secondly , we discuss the effects of temperature in a non - unital model which , in the markovian limit , gives the standard markovian master equation for a two - level atom interacting with a thermal bath . finally , we show that , as one might expect , the occurrence of non - markovianity now has a more complicated origin being linked to both the dephasing and the dissipative terms . the paper is structured as follows . in sec . ii we introduce the general time - local master equation , present its solution , and show that the noise additivity property holds beyond the markovian approximation . in sec . iii we study the complete positivity conditions , while in sec . iv we introduce a heuristic master equation which is always physical and discuss the interplay between temperature effects and non - markovianity . finally in sec . v we present conclusions . let us consider the following time - local master equation for the qubit density matrix @xmath0 in the interaction picture and in units of @xmath1 , @xmath2 + \frac{\gamma_1 ( t)}{2 } l_1 ( \rho ) + \frac{\gamma_2 ( t)}{2 } l_2(\rho ) \nonumber \\ & + & \frac{\gamma_3(t)}{2 } l_3(\rho ) , \label{eq : me}\end{aligned}\ ] ] where @xmath3 are time - dependent rates , @xmath4 is a time - dependent frequency shift , and the dissipators @xmath5 are defined as @xmath6 in the equations above @xmath7 are the inversion operators and @xmath8 ( @xmath9 ) are the pauli operators . the three dissipators @xmath10 , @xmath11 , @xmath12 , describe heating , dissipation and dephasing , respectively . however , contrarily to the typical gksl master equation @xcite , the decay rates are not positive constants but time - dependent functions which need not be positive at all times . the master equation describes phase covariant noise and has been considered recently in the context of quantum metrology in noisy channels @xcite . special cases of master equations of the form of eq . ( [ eq : me ] ) are those considered , e.g. , in refs . @xcite , for @xmath13 describing an amplitude damping model , and the pure dephasing master equation considered , e.g. , in refs . @xcite for @xmath14 . these two special cases can be derived by means of an exact approach starting from a microscopic hamiltonian model for system and environment . hence the resulting dynamics is always cptp . in the more general case considered in this paper , however , the master equation is introduced phenomenologically , since an exact microscopic derivation is unfeasible . as a consequence , restrictions on the form of the time - dependent decay rates arise in order to preserve the cptp character of the dynamics . the master equation is one of the most general time - local master equations for a qubit . indeed , it combines the effects of pure dephasing terms and dissipative terms . the dynamics is non - unital and the heating term @xmath10 accounts for the presence of a finite temperature environment . the corresponding dynamical map can be written as @xmath15 , with @xmath16 the chronological ordering operator . whenever one of the time - dependent rates takes negative values then the dynamical map is not cp - divisible , i.e. , the propagator @xmath17 defined by @xmath18 , with @xmath19 , is not cp . in the following we define as markovian a dynamics such that @xmath17 is cp @xmath20 . let us indicate with @xmath21 and @xmath22 the ground and excited states of the qubit , respectively . from eq . one straightforwardly derives the following equations for the ground state probability @xmath23 and the coherence @xmath24 : @xmath25.\end{aligned}\ ] ] the equations above are linear first - order differential equations and can be solved for any values of the time - dependent decay rates . the solution reads as follows : @xmath26 , \label{eq : p1}\\ \alpha ( t)&= & \alpha(0 ) e^{i\omega(t ) - \gamma(t)/2 - \tilde{\gamma}(t ) } , \label{eq : alfa } \end{aligned}\ ] ] where @xmath27/2 , \label{eq : gamma } \\ \tilde{\gamma}(t ) & = & \int_0^t dt ' \gamma_3(t ' ) , \label{eq : gammat}\\ \omega(t)&= & \int_0^t dt ' \omega(t ' ) , \label{eq : omega } \\ g(t)&= & \int_0^t dt ' e^{\gamma(t ' ) } \gamma_2(t ' ) /2 . \label{eq : g } \end{aligned}\ ] ] if the time - dependent coefficients quickly attain a stationary positive constant value , after an initial short time interval @xmath28 , known as the correlation time of the environment , one obtains the approximated gksl master equation by coarse - graining over @xmath28 and extending to infinity the limit of integration in eqs . - . more precisely one obtains the following markovian limits for the quantities defined in eqs . - . @xmath29 using these expressions one can recover the well known markovian formulas for populations and coherences @xmath30 the approximated gksl master equation , obtained from eq . by simply replacing the time - dependent coefficients with the corresponding positive constants , has been investigated , e.g. , in ref . @xcite to study additivity of noise in the markovian limit and for @xmath31 . there the authors show that , while for a single qubit additivity holds , composite systems may violate this property . in the single qubit case , additivity simply means that the decay rates of the off - diagonal elements of the density matrix , when the qubit is subjected to independent sources of noise , is just the sum of the decay rates arising from the interaction with each individual environment . this is straightforwardly seen in eq . . in this paper we generalise the results presented in ref . @xcite to the case of general temperatures and beyond the markovian approximation . equation , indeed , straightforwardly proves that additivity holds for the general time - local master equation , provided that the solution is physical . in the following section we will thoroughly investigate the conditions under which the master equation gives rise to a physically admittable dynamics described by a cptp map and we will give examples of both physical and unphysical behaviour for specific choices of the time - dependent decay rates . let us begin by expressing the solution in terms of components of the bloch vector defined by @xmath32 , with @xmath33 the identity operator , @xmath34 the pauli operators vector having as components @xmath8 , ( @xmath35 ) , and @xmath36 the bloch vector . the evolution of the latter one is given by @xmath37 where @xmath38 is known as the damping matrix and @xmath39 is the translation vector given by @xmath40 - 1 . \label{eq : t3 } \end{aligned}\ ] ] the eigenvalues of the damping matrix can be written as @xmath41 complete positivity conditions can be expressed in terms of inequalities involving the bloch vector components @xcite . in the following we will use the formulation introduced in ref . @xcite @xmath42 where @xmath43 , \\ q(t)=\frac{1}{2}[t_3(t)-\lambda_3(t ) ] , \\ w(t)=\frac{1}{2}[\lambda_1(t)+\lambda_2(t ) ] , \\ y(t)=\frac{1}{2}[\lambda_1(t)-\lambda_2(t ) ] . \end{aligned}\ ] ] using the analytical expressions given by eqs . - , the cp necessary and sufficient conditions read as follows @xmath44\!\ ! , \nonumber \\ \label{ycond } \\ iv ) & & e^{-\gamma(t)-2\tilde{\gamma}(t ) } \ ! \cos^2 \omega(t)\ ! \leq \ ! e^{-\gamma(t ) } \ ! \left [ \!1 - e^{-\gamma(t)}g(t ) \right ] [ g(t ) + 1 ] . \nonumber \\\label{wcond } \end{aligned}\ ] ] we notice that the validity of conditions @xmath45 ( positivity conditions ) implies that condition @xmath46 is always satisfied , as the l.h.s . of the inequality is always non - positive and the r.h.s . is always non - negative . we also stress that the dephasing term described by @xmath12 directly influences only conditions @xmath47 via the decoherence term @xmath48 . finally we note that @xmath49 and @xmath50 are sufficient conditions for positivity , i.e. , for @xmath45 . let us focus on the case in which the purely dephasing term vanishes , namely @xmath51 . in this case one sees that condition @xmath52 simplifies and can be recast as follows @xmath53 . \label{eq : ineq4}\ ] ] for @xmath54 one sees immediately that the inequality above is automatically satisfied whenever the positivity conditions @xmath45 are satisfied . in the more general case in which @xmath55 , the condition is still valid provided the positivity conditions hold since , at any time , the left hand side of the inequality is upper bounded by @xmath56 . from the reasoning above one can reach a simple conclusion regarding the physicality of the general form of master equation . indeed in this case , assuming that conditions @xmath45 are verified , a sufficient condition for complete positivity is that the term @xmath57 . we conclude this section by looking at the weak coupling and short time limits . in the weak coupling limit the following approximations hold @xmath58 using these approximations a straightforward calculation shows that conditions @xmath45 and @xmath52 correspond to @xmath59 as for the short - time approximation , by considering the taylor expansion around @xmath60 of the exponential terms , i.e. , @xmath56 and of the term @xmath61 , it is easy to convince oneself that the cp conditions amount at @xmath62 , @xmath63 , @xmath64 . these conditions imply that , contrarily to what happens in certain unital time - local master equations ( see , e.g. , ref . @xcite ) , in our model the decay rates can not take at the initial time negative values . let us now consider the following heuristic model . we assume that the open quantum system of interest is coupled to both a thermal reservoir and a dephasing environment at the same temperature @xmath16 . the former one induces heating and dissipation at rates given by @xmath65 and @xmath66 , with @xmath67 the mean number of excitations in the modes of the thermal environment . we notice that , for a zero @xmath16 environment , the heating rate @xmath68 while the dissipation rate @xmath69 . hence , we consider as a possible physically reasonable choice for the time - dependent function @xmath70 the one obtained in the exactly solvable zero-@xmath16 model presented , e.g. , in ref . @xcite . in this model the function @xmath70 takes the form @xmath71 with @xmath72 c(0),\ ] ] where @xmath73 , and @xmath74 is a dimensionless positive number measuring the overall coupling between the two - state system and the environment with respect to the width of the spectral density of the environment . the coefficient @xmath75 can be analytically calculated and yields the simple expression @xmath76 \nonumber \\ & \equiv & -\ln [ x(t)^{2n+1 } ] , \label{eq : example}\end{aligned}\ ] ] where we have used eq . and defined @xmath77 ^ 2 $ ] . we note that @xmath78 and that @xmath79 presents oscillations in time only for @xmath80 ( strong coupling , broad spectral density ) while it decays monotonically for @xmath81 ( weak coupling , narrow spectral density ) . it is straightforward to see by explicitly calculating the decay rates @xmath82 and @xmath83 that they are always positive whenever @xmath84 ( divisible dynamics ) and attain temporarily negative values for @xmath85 ( non divisible dynamics ) . , and ( b ) in the strong coupling ( non - markovian ) regime , @xmath86 , for different values of @xmath67 ( temperature).,title="fig : " ] , and ( b ) in the strong coupling ( non - markovian ) regime , @xmath86 , for different values of @xmath67 ( temperature).,title="fig : " ] by inserting eq . one obtaines the following analytic expression for the ground state population @xmath87 it is straightforward to verify that , for this model , the positivity conditions @xmath45 are verified at all times and for all values of @xmath88 . we notice that , in absence of the pure dephasing term , i.e. , whenever @xmath51 , condition @xmath52 is automatically satisfied and the corresponding @xmath16-temperature master equation is always physical . moreover , not only this model by construction reduces to the exact zero-@xmath16 model , but it also gives the correct markovian limit for a two - level system in a thermal bath at @xmath16 temperature . indeed , if we indicate with @xmath89 the markovian limit of @xmath83 in the exact zero-@xmath16 model , one can easily see that , the markovian expressions of the decoherence factor @xmath56 and of the ground state probability @xmath90 , obtained for @xmath91 , read @xmath92 , \nonumber\end{aligned}\ ] ] respectively . we now go back to the situation in which the pure dephasing term is present in the liouvillian . the coefficient @xmath93 does not influence the behaviour of the populations . in fig . 1 we plot the time evolution of the ground state population @xmath90 as a function of time for different temperatures , i.e. @xmath67 , in both the markovian case , fig . 1 ( a ) , and the non - markovian case , fig . 1 ( b ) . we notice that , for @xmath94 and for increasing values of temperature , the oscillations in ground state population are quickly damped , even if the dynamics continues to be non - markovian because both the @xmath82 and the @xmath83 decay rates take negative values . hence , the presence of oscillations in the ground or excited state probability is not just connected to the markovian or non - markovian character of the dynamics , as it was for the exact model of ref . @xcite , but depends also on the temperature of the environment . we now consider the effect of the pure dephasing term . as done before , we will again use a model of pure dephasing which arises from an exact microscopic description @xcite . in this case the analytic expression for the dephasing rate is given by @xmath95 where @xmath96 is the boltzmann constant and the spectral density is assumed to be of the ohmic class @xmath97 with @xmath98 an overall coupling constant and @xmath99 the cutoff frequency . it is worth stressing that this model always leads to @xmath100 , hence the dynamics is not only positive but also completely positive since condition @xmath52 is verified at all times . in ref . @xcite the non - markovianity of this model was studied in detail and was found to be linked to the value of the ohmicity parameter @xmath101 . hence , the two parameters governing the markovian to non - markovian crossover are @xmath74 and @xmath101 . in other words , the dynamics of the whole system can be non - markovian also for values of @xmath102 provided that the ohmicity parameter is such that @xmath103 for certain time intervals . generally , the dynamics of the coherences , given by eq . , will be damped because of both the heating / dissipation terms and the dephasing term . we notice however , that some of the characteristic phenomena typical of pure dephasing in ohmic - like environments , e.g. , coherence trapping @xcite , will not occur in this model because the coefficient @xmath104 will always eventually erase the coherences and drive the system towards a thermal mixed state . we have solved and investigated the dynamics of a general time - local master equation which combines dissipative and pure dephasing terms showing that decoherence is still additive . guided from the knowledge of the exact microscopic amplitude damping and pure dephasing master equations , we have introduced an intuitive heuristic model which is always cptp . this model allows us to study the effects of finite temperature on the dynamics of the qubit in the non - markovian regime . as expected , when increasing the temperature of the environment , decay of both populations and coherences is faster and faster . moreover , thermalisation destroys phenomena such as coherence trapping which are present in purely dephasing systems . we have pointed out that the conditions for non - markovian dynamics are now dependent on both the characteristic parameter of the dissipative terms , @xmath74 , and on the corresponding parameter for the pure dephasing term , @xmath101 . finally we have seen that finite temperature effects quickly destroy the oscillatory behaviour of populations even in the strongly non - markovian regime @xmath94 . given the importance of studies of fundamental non - markovian models , we believe that our results will be of use for both reservoir engineering and to model noise in solid - state devices in realistic experimental conditions , i.e. , when finite - temperature effects can not be neglected . as an example , an interesting future direction is the investigation of whether and how memory effects may affect the break down of additivity property in bipartite systems , as it happens in the markovian case @xcite . this work was supported by the eu collaborative project quprocs ( grant agreement 641277 ) , the academy of finland ( project no . 287750 ) and the magnus ehrnrooth foundation . d. braun , phys . lett . * 89 * , 277901 , ( 2005 ) . t. zell , f. queisser , and r. klesse , phys . rev . lett . * 102 * , 160501 ( 2009 ) . j. t. barreiro _ et al . _ , nat * 6 * , 943 ( 2010 ) . l. mazzola , s. maniscalco , j. piilo , and k .- a . suominen , phys . a * 79 * 042302 ( 2009 ) . l. viola and s. lloyd , phys . a * 58 * 2733 ( 1998 ) . l. viola , e. knill and s. lloyd , phys . lett . * 82 * 2417 ( 1998 ) . a. g. kofman and g. kurizki g , phys . lett . * 87 * 270405 ( 2001 ) . a. g. kofman and g. kurizki , phys . lett . * 93 * 130406 ( 2004 ) . j. luczka physica a * 167 * , 919 ( 1990 ) . g. m. palma , k .- a . suominen , and a .- k . ekert , proc . a * 452 * , 567 ( 1996 ) . j. q. reinga , l. quiroga , and n. f. johnson phys . a. * 65 * 032326 ( 2002 ) .	we present a general model of qubit dynamics which entails pure dephasing and dissipative time - local master equations . this allows us to describe the combined effect of thermalisation and dephasing beyond the usual markovian approximation . we investigate the complete positivity conditions and introduce a heuristic model that is always physical and provides the correct markovian limit . we study the effects of temperature on the non - markovian behaviour of the system and show that the noise additivity property discussed by yu and eberly in ref . @xcite holds beyond the markovian limit .
a hallmark of the glass transition of supercooled liquids is the time dependence of physical properties also called relaxation . this time dependence is typically seen on the timescale of the so - called maxwell relaxation time @xmath3 . @xmath3 is defined as @xmath4 where @xmath5 is the low frequency limiting ( dc ) shear viscosity and @xmath6 the high frequency limiting elastic shear modulus . the relaxation is often most easily studied in the frequency domain . at high frequencies the liquid shows solid - like elastic behaviour whereas at low frequencies it shows liquid - like viscous behaviour . in general the viscoelastic stress response to oscillatoric shear strain is described by the complex dynamic shear modulus @xmath7 @xcite . the structural relaxation is also observed in the specific heat . this has long ago been studied in the time domain as enthalpy relaxation @xcite , but in 1985 frequency domain studies or ac - calorimetry @xcite were introduced in the research field of supercooled liquids . common to these methods is that they excite thermal oscillations in the sample at a cyclic frequency @xmath8 . for a heat diffusivity , @xmath9 , this frequency is associated with a characteristic heat diffusion length , @xmath10 the magnitude of which gives the range of the excited thermal waves . when comparing @xmath11 to a given sample size , @xmath12 , one can classify ac - calorimetry experiments as to whether the sample is thermally thin ( @xmath13 ) or thermally thick ( @xmath14 ) . a typical value of @xmath9 is @xmath15 giving @xmath16 for a frequency of @xmath17 . this means that thermally thin methods are restricted to rather low frequencies @xcite or very thin samples @xcite . in these methods the temperature field is homogeneous throughout the sample and the frequency dependent specific heat is derived directly . the ac - technique allows for easy corrections of the influence of heat leaks in a non - adiabatic configuration . ac - calorimetry on thermally thick samples on the other hand has the advantage of covering a large frequency span @xcite . in this case the temperature field is inhomogenous and what is really measured is the effusivity , @xmath18 , from which the specific heat is derived . the effusivity expresses the ability of the liquid to take up heat from parts of its surface and transport it away ( thermal effusion ) . this ability is dependent on two constitutive properties , the heat conductivity , @xmath19 , and the specific heat ( pr . volume ) , @xmath20 . the specific heat can only be found to within a proportionality constant in effusion experiments from an ( infinite ) plane . however taking boundary effects into account for finite size plane heaters @xcite one may in principle find @xmath19 and @xmath20 independently . the intermediate situation where the heat diffusion length is comparable to the sample thickness has been exploited in the so - called two - channel ac calorimeter @xcite . this fine method aims directly at an independent determination of @xmath19 and @xmath20 . the specific heat entering the effusivity is widely taken as the isobaric specific heat , @xmath21 . however recently it has been shown @xcite that in planar and spherical geometries it is rather the longitudinal specific heat , @xmath22 , that enters . when @xmath23 and @xmath24 are the adiabatic and isothermal longitudinal moduli respectively , @xmath22 is related to the isochoric specific heat , @xmath25 , by @xmath26 in contrast , when @xmath27 and @xmath28 are the adiabatic and isothermal bulk moduli respectively , @xmath21 is given by @xmath29 now @xmath30 , and @xmath31 . this means that @xmath22 will differ from @xmath21 in the relaxation regime where the shear modulus , @xmath7 , dynamically is different from zero , and @xmath32 and @xmath33 differ significantly . in this work we study the frequency dependent specific heat from thermal effusion in a spherical geometry . specific heat measurements in a spherical geometry have earlier been conducted in the thermal thin limit @xcite , but in this paper we explore the thermal thick limit . the justification is threefold 1 ) we want to realize an experimental geometry in which the thermomechanical influence on the thermal effusion can be analytically calculated @xcite . 2 ) the spherical container of the liquid is a piezoelectric spherical shell by which it is possible to measure the adiabatic bulk modulus @xcite on the very same sample under identical conditions . 3 ) the heat conductivity and specific heat can be found independently , as the introduction of a finite length scale allows for separation of the two . we also in this paper give an account of the @xmath2 detection technique of our variant . the experimental method used in this study is an effusion method . in such methods a harmonic varying heat current , @xmath34 , with angular frequency @xmath8 and complex amplitude @xmath35 , is produced at a surface in contact with the liquid ; this heat then `` effuses '' into the liquid . at the surface a corresponding harmonic temperature oscillation , @xmath36 , with complex amplitude @xmath37 , is created . the current and temperature can be thought of as stimuli and response ( but which is which can be interchanged at will ) , and the two are linearly related if the amplitudes , @xmath38 and @xmath39 , are small enough . hence it is convenient to introduce the complex frequency - dependent thermal impedance of the liquid @xmath40 this impedance depends on the properties of the liquid . the terminology thermal _ impedance _ derives naturally from the close analogy between the flow of thermal heat and the flow of electricity @xcite . temperature corresponds to electric potential and heat current to electrical current . the analogy also applies to the heat conductivity which corresponds to the specific electrical conductivity , and the specific heat which corresponds to electrical capacity . as an example the thermal impedance of a thermally thin sample is given as @xmath41 where @xmath42 is the volume of the sample . again an equation completely equivalent to the electrical correspondence between capacitance and impedance . in the classical case of a planar heater in a thermal thick situation the thermal impedance is found to be @xcite @xmath43 where @xmath44 is the plate area . the lateral dimension @xmath45 has to be large ( @xmath46 @xcite . thermal effusion is possible in a spherical geometry as well as in a planar geometry . imagine a liquid confined between the to radii , @xmath47 and @xmath48 . the imposed oscillating outward heat current , @xmath49 , at radius @xmath47 results in a temperature oscillation , @xmath50 , at the same radius . the heat wave propagating out into the liquid gives rise to a coupled strain wave due to thermal expansion . the heat diffusion is thus dependent on the mechanical boundary conditions and mechanical properties of the liquid . it is widely assumed that the transport of heat in a continuum is described by the heat diffusion equation @xmath51 by the same token one might take the heat diffusion constant as @xmath52 if the surfaces at @xmath47 and @xmath48 are clamped and @xmath53 if one of the surfaces is free to move . this common belief is not entirely correct . if shear modulus is nonvanishing compared to bulk modulus the description of heat diffusion can not be decoupled to a bare heat diffusion equation @xcite . such a situation arises dynamically at frequencies where @xmath54 . the fully coupled thermomechanical equations have been discussed in detail , by some of us , in planar geometry @xcite and in spherical geometry @xcite . in general the solutions are complicated but in the case of a spherical geometry where the outer radius @xmath48 is much larger than the thermal diffusion length , @xmath55 the solution becomes simple . in terms of the liquid thermal impedance , @xmath56 , at the inner radius , @xmath47 , one has @xmath57 it is very interesting that in this thermally thick limit , the thermal impedance is independent of the mechanical boundary conditions @xcite . i.e. in the two cases of clamped or free boundary conditions mentioned before it is neither the isochoric nor the isobaric specific heat that enters the diffusion constant and the thermal impedance but the longitudinal specific heat . since a closed spherical surface has no boundary there are no correction terms like the ones to eq . ( [ eq : zplane ] ) discussed above for planar plate effusion . notice from eq . ( [ eq : tesg2 ] ) that associated with a given radius , @xmath47 , is a characteristic heat diffusion time for the liquid @xmath58 @xmath22 and hence @xmath59 may in general be complex , but when @xmath22 is real , the imaginary part of @xmath60 peaks at @xmath61 . the @xmath62 detection technique we use ( where the subscript @xmath63 indicates that it is the electrical angular frequency not the thermal ) is different from other @xmath62 techniques ( e.g. ref.@xcite ) in two ways . firstly , we include higher order terms in the temperature dependence of the thermistor used and in the power oscillations produced . secondly , we use a direct measurement of the time dependent voltage over a voltage divider instead of a lock - in amplifier . the @xmath62-detection technique makes it possible to measure the temperature with the same electrical resistor which generates the heat and thereby know the thermal current and the temperature at the same surface . the principle is in short the following : applying an electrical current at cyclic frequency @xmath64 through a resistor a combined constant ( zero frequency ) and @xmath65 component joule heating is produced . this gives rise to a combined temperature dc - offset and temperature oscillation at @xmath65 , the size of which depends on the thermal environment ( the thermal impedance ) . the resistor is chosen to be temperature dependent , it is a so - called _ thermistor_. the temperature oscillations therefore creates a perturbation of the resistance at frequency @xmath65 . the voltage across the thermistor which by ohm s law is the product of current and resistance thus contains a @xmath62 ( and an additional @xmath66 ) component proportional to the temperature oscillations at @xmath65 . in practise one rather has a voltage source than a current source . in order to get a @xmath62 signal the thermistor is therefore placed in series with a temperature independent preresistor as detailed in the following . the result of the analysis is the frequency dependent thermal impedance at a frequency of @xmath65 . it is convenient to express this thermal impedance as function of the frequency of the thermal signal which we designate @xmath8 and notice that @xmath67 . we first outline the simplest theory in order to elucidate the technique most clearly . however , to increase the signal to noise ratio it is expedient to choose a thermistor with a large temperature dependence and to choose temperature amplitudes as high as possible although within the linear regime of the liquid thermal response . this necessitates higher harmonics be taken into account . this detailed analysis is deferred to appendix [ app : achigher ] . complex numbers are known to be of great use in harmonic analysis of linear systems . since the heat current is quadratic in the imposed electrical current and the whole idea of the @xmath62-detection technique is to exploit the inherent nonlinearity introduced by the temperature dependence of the thermistor the formalism becomes a little more troubled than in pure linear applications . thus we have to drag around the complex conjugate terms in order to get product terms right . generally a sum of harmonic terms , @xmath68 with real amplitudes , @xmath69,@xmath70 , and phases , @xmath71 , can be written as , @xmath72 by introducing the complex amplitudes @xmath73 and the shorthand notation @xmath74 the sum becomes @xmath75 where @xmath76 means that the complex conjugated of all the terms within the parenthesis should be added , including the real constant term . note that @xmath77 , @xmath78 , and @xmath79 . notice that any product of an @xmath80th and @xmath81th harmonic will produce both an @xmath82th and @xmath83th harmonic term . furthermore the dc - components get doubled because there is no difference between @xmath84 and @xmath85 . consider the diagram of the electrical setup seen in fig.[fig : vdiv ] . @xmath86 is the voltage of the source and @xmath87 the voltage across the temperature independent preresistor , @xmath88 . @xmath86 and @xmath87 can be measured directly . @xmath89 is the temperature dependent resistance of the heater / sensor ( we will explicitly state the temperature / time dependencies when appropriate , otherwise it is implicitly assumed ) . -detection technique . the temperature - dependent resistor ( the thermistor ) , @xmath90 , is in thermal contact with the liquid probing its thermal impedance at @xmath65 . the preresistor , @xmath88 , is temperature - independent . ] the thermistor and the surrounding sample are placed in a cryostat that defines an overall reference temperature , @xmath96 . the power produces a ( real ) temperature change @xmath97 at the heater / sensor depending on the thermal impedance , @xmath98 , ( as defined in section [ sec : linresponsetheory ] ) the heater / sensor looks into . the temperature rise has a real dc - component , @xmath99 , and complex ac - components . it can in general be written as @xmath100 the dominant terms will be the dc component , @xmath99 , and the second harmonic term , @xmath101 ( see next section ) . to first order in the temperature variations the thermistor has a resistance of @xmath102 here @xmath103 and @xmath104 are considered constant at a given cryostate temperature but they vary with @xmath96 . in fact the thermistor , which is based on a semiconducting material , has a resistance that over a larger temperature range is very well described by @xmath105 with @xmath106 the activation temperature and @xmath107 the infinity temperature limiting resistance . @xmath104 is related to @xmath106 by @xmath108 characteristic values of @xmath109 and @xmath106 are given in table [ tab : expparm ] . .experimental parameters . * top : * parameters characterizing the resistors in the setup , @xmath88 is the resistance of the preresistor , @xmath107 and @xmath106 the characteristics of the thermistor according to eq . ( [ eq : rfull ] ) ) . * middle : * absolute size of the thermistor bead . measured values indicates the eccentricity of the bead . the optimize value is the one use in the final treatment of the data , and is optimized to give best correspondence between the room temperature dc specific heat , and the literature value of this . * bottom : * applied amplitudes expressed as mean applied voltage , ( @xmath110 ) , and maximum temperature amplitude , ( @xmath111 . [ cols= " < , < " , ]	we present a novel method of measuring the frequency dependent specific heat at the glass transition applied to 5-polyphenyl-4-ether . the method employs thermal waves effusing radially out from the surface of a spherical thermistor that acts as both a heat generator and thermometer . it is a merit of the method compared to planar effusion methods that the influence of the mechanical boundary conditions are analytically known . this implies that it is the longitudinal rather than the isobaric specific heat that is measured . as another merit the thermal conductivity and specific heat can be found independently . the method has highest sensitivity at a frequency where the thermal diffusion length is comparable to the radius of the heat generator . this limits in practise the frequency range to @xmath0@xmath1 decades . an account of the @xmath2-technique used including higher order terms in the temperature dependency of the thermistor and in the power generated is furthermore given .
not much is known about the size of the population of minor bodies of the solar system between the orbits of jupiter and neptune , nor how these bodies are distributed there . the known representatives of that population , the centaurs , are the objects with @xmath3 au intermediate between comets and asteroids , having presumably short lifetimes compared to the solar system s age ( asher & steel 1993 ) . these objects , as well as the other known populations , such as the jupiter - family comets , could have a common origin in the kuiper belt ( luu & jewitt 1996 ) . for jupiter - family comets , an alternative evolutionary path from the oort cloud has been found unlikely ( see e.g. review by kresk 1994 and refs . therein ) . if the kuiper belt objects ( called ` kuiperoids ' hereinafter ) are indeed responsible for progressive replenishment of the observable populations , gravitational scattering on all four giant planets would be responsible for the transport of these objects from the trans - neptunian region all the way inward , down to jupiter ( levison & duncan , 1997 ) . the present paper aims at a related , but rather different subject , _ viz . _ computing the distributions of minor bodies between neptune and jupiter , including analysis of those distributions in the space of orbital coordinates , @xmath1 . an accompanying paper deals with the distribution of dust expected to be produced by these bodies . gravitational scattering of test bodies on the planets ( carusi , valsecchi , & greenberg , 1990 ) and influence of resonances ( jackson & zook 1989 , roques et al . 1994 ) have been studied previously by a number of investigators . here , we undertake a combined numerical study of these effects to demonstrate that , when put together , they mutually enhance each other by creating qualitatively new phenomena . while using a different numerical approach and employing different tools to analyze the computational results , we also make emphasis on a different subject the stationary distributions of comets near each of the giant planets and its structure in the space of orbital coordinates . [ a systematic use of distributions in these coordinates is a part of the kinetic approach employed by us in studying the structure , evolution , and origin of the interplanetary dust and its sources ( gorkavyi et al . 1997a , b , 1998 ) ] . furthermore , as opposed to the usual consideration of a _ dissipative _ mechanism of captures into mean motion resonances ( e.g. , due to the tidal dissipation or poynting - robertson drag ) , we deal here with an entirely different type of _ non - dissipative _ captures , when the tisserand constant is conserved . indeed , gravitational scattering as the only accompanying process in our consideration here is an elastic process . a dissipationless capture of a minor body into a mean motion resonance with the planet becomes possible at a large eccentricity of a minor body , when an exchange of angular momentum between the planet and the minor body becomes especially effective due to their very close approach . in our numerical study , we adopt the approximations of a restricted 3-body problem ( the sun , the planet on a circular orbit , and a massless minor body ) . it is convenient to define the planet s zone of gravitational influence in the @xmath4-plane of orbital coordinates : @xmath5 where @xmath6 is the semi - major axis of a test body , @xmath7 is the semi - major axis of the planet , and @xmath8 is eccentricity of the test body . this zone looks like a triangle and we call it hereinafter ` the triangle zone ' . our study examines the distribution of minor bodies within each planet s zone of influence . we also evaluate qualitatively how the bodies are scattered from one zone to another . obviously , a restricted 3-body problem is a highly simplified approach to examine the distribution of minor bodies in the gravitational field produced by the four giant planets . however , as the reader will see below , even this approximation reveals a rich and sophisticated structure in the distribution of test bodies . our follow - up research will study how the revealed distribution is modified when the influence of three other planets , with accounting for non - circular orbits as well , is taken into consideration . a quasi - stationary distribution of minor bodies has been established in the outer parts of the solar system due to multiple gravitational scatterings on the four giant planets . each giant planet maintains this quasi - stationarity in its zone of gravitational influence by ejecting the minor bodies in the amounts comparable to what flows into this zone from outside . as a convenient approach to simulate such a quasi - stationary distribution of massless minor bodies around a giant planet , we applied the following computational procedure : a record of coordinates and velocities of a test body was taken after certain number of revolutions ( usually each 10 revolutions ) of the planet around the sun and these data were then used to characterize the positions of _ many _ test bodies over the entire time span , beginning at an initial instant , and ending at the instant of ejection of the test body from the planet s zone of influence . we computed 44 stationary distributions of test bodies totalling @xmath9 @xmath10-orbital elements . details of computational runs are given in table 1 . % % put table 1 here%% meanwhile a small fraction of minor bodies is scattered from the zone of influence of each outer giant planet into the next innermost giant neighbor s zone . as a result , there is a flow of these objects from the trans - neptune region inward the solar system . our results of orbit integrations are shown , in the orbital coordinates @xmath11 and @xmath12 , for neptune in fig . 1a , b ; for uranus in fig . 3a , b ; for saturn in fig . 5a , b ; and for jupiter in fig . 7a , b . the minor bodies ( kuiperoids and centaurs ) from the trans - neptune regions can be captured into mean motion resonances with neptune ( morbidelli 1997 , levison & dunkan 1997 , malhotra 1998 ) . a minor body being in a resonance with neptune increases its eccentricity , approaches closely the planet and undergoes a strong gravitational scattering on it . the scattering results in filling a quasi - triangle zone with numerous resonances and gaps ( see fig . the minor bodies populating neptune s zone of gravitational influence should form a broad belt along neptune s orbit the ` neptune belt ' . furthermore , a fraction of minor bodies from neptune s quasi - triangle zone is intercepted by uranus and fills in the quasi - triangle zone of the latter in the same fashion as it happens for neptune s zone of influence . a similar process of ` leakage ' of a part of minor bodies occurs from the belt of each outer giant planet into the next innermost giant neighbor s sphere of influence . some of them ( e.g. uranus ) are able to eject , by gravitational scattering , a minor body with such a high velocity in a particular direction that the body can reach jupiter . as a net result , there is a flow of these objects from the trans - neptune region downward to jupiter . the minor bodies form a rather wide belt around each giant planet s orbit . 1a , 3a , 5a , and 7a indicate that the bulk of test bodies is located within each planet s triangle zone of influence , although a substantial number of bodies form a ` beard - like ' distribution with lower eccentricities . two types of structures in the zone of gravitational influence of each giant planet are clearly seen in figs . 1 to 8 : ( i ) resonant gaps and ( ii ) resonant groups of test bodies . there is a substantial difference between them : gaps represent a robust feature , whereas probability to be captured into a resonant group is rather sensitive to the number of computational runs used . we would like to emphasize an interesting phenomenon of a _ near - resonance accumulation _ : the test bodies tend to avoid location directly within the resonances , both interior and exterior ones . instead , they prefer to be positioned slightly aside , mostly at both sides of the resonant gaps . those gaps seem to be similar to the well - known kirkwood gaps in the main asteroid belt . the locations of gaps are clearly seen from the number of test bodies as a function of semi - major axis , which is shown , for each of the giant planets , in figs . 2 , 4 , 6 , and 8 occurence of locations is a measure of a probability for a test body , @xmath13 , to be located between @xmath6 and @xmath14 ( with @xmath6 measured in @xmath15 , where @xmath16 and the total probability to be in the planet s triangle zone is @xmath17 . there are two possible mechanisms for gap formation . the first one is associated directly with gravitational scattering of a test body on the planet . the scattering can be stronger when the minor body s period of revolutions is commensurable with the planet s period , and this is not accompanied by the resonant capture thereby producing a gap . the second mechanism is associated with the resonance influence : by getting an excessive angular momentum from the planet , the minor body leaves the resonance . when the minor body leaves the resonance , the balance of the angular momentum forces the minor body to librate , thereby spending much of the time at the edge of the resonant region . this results in a _ near - resonant accumulation_. we observe numerous examples of such accumulations ( see e.g. figs . 3 and 5 ) . without going into a detailed classification , we will not discriminate here between usual resonance and near - resonant accumulation . test bodies captured into resonances in the triangle zone , undergo a characteristic pattern of evolution in @xmath8 and @xmath18 within the resonances : as long as the eccentricity of the minor body captured into a particular resonance increases ( decreases ) , its inclination ( shown in figs . 1b , 3b , 5b , and 7b ) decreases ( increases ) so that the tisserand constant , @xmath19 , is kept invariable . in our computations , @xmath19 is only insignificantly ( at the typical level of a fraction of 1% ) fluctuates . a detailed consideration of the dynamics leading to the capture into , as well as strength of , resonances when gravitational scattering is the only accompanying process is considered elsewhere ( gorkavyi & ozernoy 1999 ) . for illustrative purposes , in fig . 9 we show in @xmath4-coordinates the positions of all four cometary belts we expect to exist near the orbits of the four giant planets . the known minor bodies in the main asteroid belts and the known centaurs are also shown fig . although this picture neglects eccentricities of the giant planets as well as gravitational perturbations from three other giant planets in each planetary zone of influence ( thereby secular resonances are out of consideration here ) , the resonant patterns in each zone are expected to be preserved . it is remarkable that , as is seen in fig . 9 , at least 4 of 6 known centaurs are located close to the resonances . finally , figs . 10 and 11 illustrate , in the above approximations , the large - scale structure of the solar system formed by the cometary - asteroidal belts envisaged in this paper . the maxima in minor body concentrations clearly delineate the positions of these belts , although the actual density contrasts should be determined in further , more detailed computations . as our simulations indicate , the distribution of cometary and asteroidal bodies between the orbits of jupiter and neptune is expected to be structured into four belts . these belts are well separated in the @xmath20)-plane of the orbital coordinates . moreover , they are expected to be distinguishable in space as well . the cometary - asteroidal belts near all four giant planets contain a rich resonant structure . each belt plotted in the orbital coordinate @xmath10-space , is characterized by sharp edges and an internal structure that includes density maxima and gaps around resonances . the latter are similar to the kirkwood gaps in the main asteroid belt . the minor bodies near the boundary of ( or inside ) the planet s triangle zone described by eqs.(1)-(2 ) can be captured into non - dissipative resonances . more precisely , the bodies tend to be located not directly in the resonances but , instead , they reveal a near - resonance accumulation on the both sides of the resonance . dynamics of such bodies can be explained by a balance between resonant interactions with the planet and close approaches to it . we expect that an appreciable part of newly discovered centaurs between jupiter and neptune be in resonances with the appropriate planet and located outside the ` triangle zones ' of the giant planets . the resonant structure of the cometary - asteroidal belts near the outer planets simulated in this work has some general features which could be revealed both in the distribution of aaa - asteroids near earth and near exo - planets . the envisaged cometary - asteroidal belts , which represent the largest structures in the solar system , are expected to serve as the major sources of dust in the outer parts of our planetary system . the dust distribution produced by these sources is the subject of our accompanying paper . this work has been supported by nasa grant nag5 - 7065 to george mason university . n.g . acknowledges the nrc - nas associateship . t.t . is thankful to the american astronomical society for a small research grant from the gaposchkin s research fund . * a. eccentricity vs. the semi - major axis of minor bodies in and near neptune s zone of influence from a simulation of 29,885 test body positions . the sides of the ` triangle zone ' given by eqs . ( 1)-(2 ) are shown by dashed lines , and its base ( the upper boundary of eccentricities of test bodies ) is determined by the tisserand constant . test bodies on resonant and chaotic orbits below the triangle zone s sides delineate a ` beard ' . both the triangle zone and its ` beard ' contain numerous gaps and resonances , including those of a high order ( e.g. 7:1 , 6:1 , etc . ) b. inclination vs. the semi - major axis of minor bodies in and near neptune s zone from the same simulation . note an interesting two - side accumulation of test bodies around 4:3 resonance as well as one - side accumulation of test bodies around 1:1 resonance . b. inclination vs. the semi - major axis of minor bodies in and near uranus zone from the same simulation . note several ` muffs ' ( i.e. symmetric accumulations localized within a short range of inclinations ) near the resonances 2:1 , 3:1 , 6:1 , and 7:1 . a. eccentricity vs. the semi - major axis of minor bodies in and near saturn s zone of influence from a simulation of 28,848 test body positions . numerous gaps and are clearly seen in the resonances 7:1 , 13:2 , 6:1 , etc . b. inclination vs. the semi - major axis of minor bodies in and near saturn s zone from the same simulation . note several ` muffs ' similar to those seen in fig . they are formed at @xmath30 , i.e. near the edge of the triangle zone . . a*. eccentricity vs. the semi - major axis of minor bodies in and near jupiter s zone of influence from a simulation of 23,602 test body positions . one can see that the jovian gaps have wider and less sharp edges compared to those of less massive giant planets . . occurence of various resonances and gaps in and near jupiter s zone from the same simulation as fig . 5 . the structure of the resonant region between resonances 2:1 and 4:1 looks very similar to the kirkwood gap in the main asteroid belt between 1:4 and 1:2 resonances . . the large - scale structure of the outer part of the solar system shown in the orbital @xmath4-coordinates . the simulated cometary - asteroidal belts are shown along with the currently known objects . crosses stand for asteroids of the main belt ( 100 objects ) , triangles stand for jupiter - family comets ( 112 objects ) , squares stand for centaurs ( 6 objects ) , and diamonds stand for kuiperoids ( 50 objects ) . the simulated population fills the densest parts of the triangle zones of all the giant planets . the known centaurs and those kuiperoids whose orbits cross the neptune s orbit are in this densest part as well . one can see that bulk of the known jupiter - family comets are located far from the simulated , yet unrevealed , cometary population in the densest part of jupiter s triangle zone . we explain this by observational selection : the major part of the known jupiter - family comets has perihelia @xmath31 au ( above the dashed line ) . the sides of the ` triangle zone ' given by eqs . ( 1)-(2 ) are shown by heavy lines . . the large - scale structure of the outer part of the solar system shown face - on . for convenience , positions of jupiter , saturn , uranus , and neptune are indicated . one can see ( especially clearly for neptune and jupiter ) the simulated cometary - asteroidal belts associated with the giant planet orbits . . the large - scale structure of the outer part of the solar system shown edge - on . for convenience , positions of jupiter , saturn , uranus , and neptune are indicated . one can see ( especially clearly for neptune and jupiter ) the simulated cometary - asteroidal belts associated with the giant planet orbits . [ asher , d.j . & steel , d.i . 1993 , mon . not . 263 - 179 ] [ carusi , a. , valsecchi , g.b . & greenberg , r. , 1990 , celest . 49 , 111 ] [ gorkavyi , n.n . , ozernoy , l.m . & mather , j.c . 1997a , apj 474 , 496 ] [ gorkavyi , n.n . , ozernoy , l.m . , mather , j.c . & taidakova , t. 1997b , apj 488 , 268 ] [ gorkavyi , n.n . , ozernoy , l.m . , mather , j.c . & taidakova , t. 1998 , earth , planets and space , vol . 50 , 539 ] [ gorkavyi , n.n . , ozernoy , l.m . , 1999 , apj ( to be submitted ) ] [ jackson , a.a . & zook , h.a . 1989 , nature 337 , 629 ] [ kresk , @xmath32 . 1994 , in _ asteroids , comets , meteors 1993 _ ( a. milani et al . , , p. 77 ] [ levison , h.f . & duncan m.j . 1997 , icarus 127 , 13 ] [ luu , j.x . & jewitt , d.c . 1996 , aj 112 , 2310 ] [ malhotra , r. 1998 , in _ solar system formation and evolution " _ ( d. lazzaro et al . 149 , 37 ] [ marsden , b.g . 1998 , mpec 1998-v14 : distant minor planets ] [ morbidelli , a. 1997 , icarus 127 , 1 ]	we employ an efficient numerical approach to simulate a stationary distribution of test objects , which results from their gravitational scattering on the four giant planets , with accounting for effects of mean motion resonances . using the observed distribution of the kuiper belt objects , we reconstruct , in the space of orbital coordinates , the distribution function @xmath0 for the population of minor bodies beyond jupiter . we confirm that thousands of large yet cold comets and centaurs might be located between the orbits of jupiter and neptune . moreover , we find as an important result that they are concentrated into four circumsolar belts , with a highly non - uniform and well structured distribution of the objects . this huge yet unrevealed population , with only a few of its representatives presently known , is expected to have , like our simulations demonstrate , a rich resonant structure containing both density maxima and gaps . the resonant structure is formed due to gravitational perturbations , i.e. in a non - dissipative way . if plotted in the ( @xmath1)-space of orbital coordinates , the belts contain gaps ( including those between resonant groups ) , quite similar to the kirkwood gaps in the main asteroid belt . an appreciable fraction of the test bodies reveals , for some time , an accumulation near ( rather than in ) the resonances , both interior and exterior , with the giant planets . an accompanying paper considers the population simulated in this work as the major source of dust in the outer solar system . the simple but fast and efficient numerical approach employed in this work would allow the reader for applying it to many other problems of his / her interest . @=11 # 1 # 1 @=12 # 1@xmath2 # 1#2to 6.5truein 9.25truein -1 cm
galactic bars are frequently invoked as candidates for facilitating the transfer of mass from the interstellar medium of seyfert galaxies to their central engines ( e.g. schwartz 1981 ; norman 1987 ; shlosman , frank & begelman 1989 ) . however , the importance of bars in seyfert galaxies remains controversial , with some studies suggesting that seyfert nuclei occur preferentially in barred systems ( e.g. arsenault 1989 ) , while others find no such preference ( heckman 1980 ; simkin , su , & schwarz 1980 ; moles et al . most comparisons of the hosts of normal and active galaxies have been made at optical wavelengths , where the effects of dust and star formation can mask bar structures . near - infrared images provide a better tracer of the mass distribution including bars since the relative importance of these effects is reduced at near - infrared wavelengths . several k - band imaging studies of seyfert galaxies have revealed the presence of bar structures in galaxies classified as unbarred in the optical ( e.g. scoville et al . 1988 ; thronson et al . 1989 ; mcleod & reike 1995 ) , demonstrating the benefit of working in the near - infrared . in this _ letter _ , we use a recently completed modified k - band imaging survey of samples of seyfert and normal galaxies ( mulchaey , regan & kundu 1997 ; hereafter paper i ) to study the role of bars in the fueling of nuclear activity . the seyfert sample was selected from the revised shapely - ames ( rsa ) and the sandage & tammann ( 1987 ) extension to the rsa . all of the galaxies observed have recessional velocities less than 5000 km s@xmath2 and logarithmic axial ratios ( log a / b ) less than 0.2 . the axial ratio limit was adopted to avoid highly inclined galaxies where bars can be difficult to recognize . the control sample of normal galaxies was also selected from the rsa and was matched to the seyfert sample in hubble type , redshift , inclination and blue luminosity . a total of 30 seyfert and 25 normal galaxies were observed in the near - infrared with a modified k - band filter . a complete description of the sample properties and observing conditions are given in paper i. the bar morphology of each galaxy was determined using the criteria outlined in detail in paper i. approximately 50% of the galaxies in both samples are classified as barred in the rsa catalog . we detect near - infrared bars in all of these galaxies . we also find evidence for bars in sixteen of the twenty - nine galaxies classified as unbarred in the rsa . this further demonstrates the gains that can be made by observing galaxies in the near - infrared . overall , @xmath0 70% of the galaxies observed show evidence for a bar at 2.1@xmath1 m . figure 1 summarizes the bar statistics in paper i. we note that the percentage of barred systems is comparable in the seyfert and normal galaxy samples . this suggests that seyfert nuclei do not occur preferentially in barred galaxies , a result consistent with earlier studies at optical wavelengths ( heckman 1980 ; simkin , su , & schwarz 1980 ; moles et al . . a slightly higher percentage of the seyfert 2 galaxies than seyfert 1 galaxies appear to be barred in our images ( 83% vs. 67% , respectively ) , but this result is not statistically significant given the small number of objects of each type . while we find no significant difference between the percentage of barred galaxies in our normal and active galaxy samples , there may still be differences between the seyfert hosts and spiral galaxies in general . for example , in selecting our control sample , we have matched the normal galaxies to the seyfert galaxies in total blue luminosity . however , the seyfert host galaxies tend to be more luminous on average than the typical spiral galaxy in the rsa catalog . thus , our control galaxies are more luminous in the blue than typical rsa spirals and our results for the control sample may not reflect the characteristics of typical spirals . in fact , we note that the percentage of bars in both the seyfert and control sample is higher in the rsa ( i.e. 42% and 45% ) than the overall percentage of barred galaxies in the rsa for similar hubble types ( for example , only 25% of the rsa spirals of type sa or sab are barred ) . therefore there may be a trend between the luminosity of the host and the presence of a bar . while ideally we would like to make a direct comparison between the bar properties of our seyfert galaxies and the rsa spirals in general , near - infrared images for large samples of spirals over a range in luminosities do not yet exist . from the present data , we can conclude that the percentage of barred galaxies is comparable among seyfert galaxies and normal spiral galaxies of similar luminosity . shlosman , frank & begelman ( 1989 ) have proposed a bars within bars mechanism that would drive material collected from a large - scale bar down to the scales where the central supermassive blackhole dominates the gravitational potential ( @xmath0 inner 10 pc ) . in their model , gas driven inward from the large - scale stellar bar accumulates in the central few hundred parsecs in a rapidly rotating disk . if the mass of the gas in this disk is an appreciable fraction ( @xmath3 20@xmath4 ) of the dynamical mass at that radius , the disk may become unstable and a gaseous bar can form . a dynamically unstable series of such bars may exist , continuing the inflow of material towards the nucleus . nested bars have been observed in some galaxies ( e.g. shaw et al . 1995 ; wozniak et al . 1995 ; friedli et al . 1996 ) , possibly supporting the above scenario . in fact , wozniak et al . ( 1995 ) have noted a high percentage of seyfert galaxies in their sample of double - barred systems . we have searched for the presence of multiple bar structures in our near - infrared images . figure 1 shows , however , that the percentage of double - barred galaxies is actually higher in our normal galaxy sample than in our seyfert sample . the failure to find multiple bars in most of our galaxies could be a result of our limited resolution . at the typical distance of our sample galaxies , our spatial resolution is @xmath0 300500 pc ( h@xmath5 = 50 km s@xmath2 mpc@xmath2 assumed ) . higher resolution images may uncover many more examples of multiple bars . we also note that the bars within bars mechanism proposed by shlosman et al . ( 1989 ) relies on _ gaseous _ bars in the inner kiloparsec and these bars may not always be accompanied by the stellar bars that our images probe . high resolution observations of atomic and molecular gas in seyferts may be required to fully test the shlosman et al . ( 1989 ) scenario . piner , stone & teuben ( 1995 ) have recently suggested that bars with a high axial ratio ( i.e. thin bars ) can also move mass efficiently into the inner 100 pc of a galaxy . a comparison of the maximum ellipticity for the bars in the two samples ( figure 2 ) suggests no differences in the bar axial ratios of the seyfert and normal galaxies . furthermore , the two galaxies with the highest bar axial ratios are normal galaxies . thus , we find no evidence that seyfert galaxies preferentially contain thin bars . while bars appear to be a common feature in seyfert galaxies , @xmath0 30% of our sample galaxies show no evidence for a bar even in the near - infrared . it seems likely that most of these galaxies do not contain bars . a particularly striking example is the seyfert 1 galaxy ngc 7213 , which has nearly circular isophotes from the center down to a surface brightness level of @xmath0 19 k mag arcsecond@xmath6 . the failure to find bars in even this small percentage of seyferts indicates that bars are probably not a ubiquitous feature of active galaxy hosts . mcleod & reike ( 1995 ) reached a similar conclusion from their k - band study of the cfa seyfert sample . a relevant question is whether the non - barred seyferts show evidence for other perturbations that might be responsible for fueling the nuclear activity . galaxy interactions and encounters have often been suggested as a mechanism to induce activity in galaxies and in fact several of the non - barred seyferts appear to be interacting with neighboring galaxies ( e.g. ngc 5427 ; kennicutt & keel 1984 ) . however , in other cases , the nearest neighbor is much too far away for gravitational interactions to be a reasonable scenario ( e.g. ngc 788 nearest neighbor is over 1 h@xmath7 mpc away ; moles et al . 1995 ) . in addition , when a major merger interaction drives gas toward the center of a galaxy it does so by forming a bar ( mihos & hernquist 1996 ) . since these possibly interacting seyferts are non - barred , it is not clear how the interaction could be transporting mass to the central engine . moles et al . ( 1995 ) have suggested that although bars may not be more common in seyfert galaxies than normal spirals , other features often associated with bars , such as rings , may be more prevalent in galaxies with nuclear activity . however , rings are a secondary probe of the potential and not as straightforward a probe as our near - infrared images . thus , a non - detection of a bar in the near - infrared probably rules out the presence of a strong nonaxisymmetric component to the potential even when the galaxy has a ring . it has been proposed that the formation of a black hole at the center of a galaxy can lead to the destruction of a bar ( norman et al . 1996 ; friedli & benz 1993 ) . the simulations of norman et al . ( 1996 ) show that the bar can be destroyed in only a fraction of the bar rotation time once the mass of the central black hole approaches 5% of the total mass of the galaxy . for most barred galaxies this means that bars can be destroyed in only @xmath0 10@xmath8 years . in this case , the stars that made up the bar are distributed into random axisymmetric orbits that give the appearance of a bulge . this very fast destruction time scale means that it may be possible for a black hole at the center of a galaxy to remain active even if the bar that supplied the fuel has been destroyed . these non - barred seyferts should exhibit certain characteristics if they are formerly barred galaxies . the stellar population of the bulge of these galaxies should be the same age as the stellar population of the disk . in addition , the central velocity dispersion of these galaxies should be quite high since the central black hole should contain @xmath05% of the total mass of the galaxy . therefore , the velocity dispersion of the centers of the non - barred seyferts should imply a higher mass fraction in the central region than the velocity dispersions of the central regions of the barred seyfert galaxies , where the ratio of black hole to galaxy mass is presumably lower . we have used a large near - infrared imaging survey to study the bar properties of seyfert and normal galaxies . we find near - infrared bars in all of the sample galaxies previously classified as barred in the revised shapely - ames catalog and in @xmath0 55% of the galaxies previously classified as unbarred . approximately 70% of the seyfert galaxies are barred , with @xmath0 10% having multiple - bar structures . the percentage of bars is comparable in the normal galaxy sample . in general , the global properties of bars , such as axial ratio , appear to be similar in seyfert and normal spiral galaxies of similar luminosity . a significant fraction of the seyferts studied show no evidence for bars even in the near - infrared . this suggests that either large - scale bars are not a universal mechanism for the transfer of mass to the central engine or that the bars in the observed non - barred seyferts were destroyed , possibly with the formation of the central black hole . figure 1fraction of barred and double - barred systems in the seyfert and control galaxy samples . the percentage of barred galaxies is comparable in the seyfert and control samples . a slightly higher percentage of the normal galaxies have double bars .	we use a recent near - infrared imaging survey of samples of seyfert and normal galaxies to study the role of bars in the fueling of nuclear activity . the active galaxy sample includes seyfert galaxies in the revised shapely - ames ( rsa ) and sandage & tammann s ( 1987 ) extension to this catalog . the normal galaxies were selected to match the seyfert sample in hubble type , redshift , inclination and blue luminosity . all the galaxies in both samples classified as barred in the rsa catalog are also barred in the near - infrared . in addition , @xmath0 55% of the galaxies classified as non - barred in the rsa show evidence for bars at 2.1 @xmath1 m . overall , @xmath0 70% of the galaxies observed show evidence for bar structures . the incidence of bars in the seyfert and normal galaxies is similar , suggesting seyfert nuclei do not occur preferentially in barred systems . furthermore , a slightly higher percentage of normal galaxies have multiple - bar structures . a significant percentage of the seyfert galaxies in our sample show no evidence for the presence of a bar even in the near - infrared . this suggests that either large - scale kiloparsec bars are not a universal fueling mechanism in seyfert galaxies or that the bars in the non - barred seyferts were recently destroyed , possibly by the formation of the central black hole . and
the foundations of analyzing thermal and vacuum fluctuations of the electromagnetic field inside matter were laid in the seminal work of s.m.rytov@xcite . this later gave rise to a unified approach of understanding fluctuational forces@xcite ( lifshitz theory of casimir forces ) , near field thermal emission and radiative heat transfer@xcite . ( polder - van - hove theory@xcite ) . recent developments in nanoengineering and detection have led to experimental regimes@xcite where these effects can play a dominant role . simultaneously , theoretical work has shed light on the fact that the classical scattering matrix along with the temperatures of objects of various geometries can completely characterize these fluctuations in both equilibrium and non - equilibrium situations@xcite . metamaterials are artificial media designed to achieve exotic electromagnetic responses that are beyond those available in conventional materials@xcite . a large body of work has emerged in the last decade which in principle engineers the classical scattering matrix to achieve effects such as negative refraction@xcite , enhanced chirality@xcite , invisibility@xcite and subwavelength imaging@xcite . recently , it was shown that a specific class of metamaterials , known as hyperbolic media@xcite(indefinite media ) has the potential for thermal engineering . such media support unique modes which can be thermally excited and detected in the near - field due to the super - planckian nature of their thermal emission spectrum@xcite . in this paper , we adopt the techniques of fluctuational electrodynamics to provide a first - principle account of the thermal emission characteristics of hyperbolic media . we show that the conventional approach of utilizing the second kind of fluctuation dissipation theorem@xcite is equivalent to the scattering matrix method@xcite for calculating the metamaterial energy density . we specifically provide the derivations of the fluctuational effects in both effective medium theory and practical thin film multilayer metamaterial designs@xcite . while the characteristics can in principle be obtained from formulas related to the reflection coefficients , it does not shed light on various aspects of equilibrium or non - equilibrium fluctuations in the context of metamaterials . our aim is to provide an insightful look at prevailing approaches adopted to the case of hyperbolic media . we also consider the case of a practical phonon - polaritonic metamaterial@xcite and show the stark contrast in the far - field and near - field thermal emission characteristics@xcite . this should help experimentalists design experiments starting from analyzing the far - field characteristics , retrieving effective medium characteristics and then look for our predicted near - field effects . we show that the far - field characteristics are dominated by the epsilon - near - zero and epsilon - near - pole responses as expected from kirchoff s laws@xcite . this is true independent of material choice and can occur for both nanowire and multilayer hyperbolic media@xcite . we comment here that for practical applications high temperature plasmonics and metamaterials would be needed@xcite . we also study the limitations of effective medium theory ( emt ) but focus on cases where there is good agreement between practical structures and emt@xcite . we emphasize that it is known in the metamaterials community that the unit cell of a metamaterial can show characteristics similar to the bulk medium@xcite . in the context of thin film hyperbolic media , this was experimentally elucidated in ref . and theoretically explained in detail in ref . . in this paper we also describe another effect connected to hyperbolic super - planckian thermal emission@xcite . we analyze the spatial coherence@xcite of the near - field thermal emission and relate it to the metamaterial modes . we show that there is a subtle interplay in near - field spatial coherence due to competition between surface waves and hyperbolic modes . we expect our work to aid experimentalists in isolating thermal effects related to metamaterials and also form the theoretical foundation for developing the macroscopic quantum electrodynamics@xcite of hyperbolic media . in global thermal equilibrium , the first kind of fluctuation dissipation theorem@xcite(fdt ) directly specifies the correlation function of electric fields . it is expressed by @xmath0 here @xmath1 is the dyadic green s function@xcite(dgf ) , @xmath2 is the mean energy of a thermal oscillator . ( [ first kind ] ) has two main applications . firstly , it can be used to derive the electromagnetic stress tensor at a certain point . secondly , it directly gives the cross - spectral density tensor@xcite which characterizes the spatial coherence of a thermal radiative source . the second kind of fdt@xcite that specifies the correlation function of thermally generated random currents is @xmath3 we assume the permittivity @xmath4 is a diagonal matrix ; @xmath5 denotes the imaginary part . the first kind of fdt can only be used in global thermal equilibrium . in non - equilibrium situation , we should first employ maxwell equations to obtain the electromagnetic fields generated by random currents through the dgf , @xmath6 and then calculate the electromagnetic stress tensor or the cross - spectral density tensor . the dyadic green s function ( dgf ) satisfies an important identity@xcite , @xmath7 this identity ensures that at global thermal equilibrium the first kind and the second kind of fdt lead to identical results . in this section , we consider an uniaxial medium located in the lower space @xmath8 at temperature t while the upper space vacuum part is at zero temperature . the relative permittivity of the uniaxial medium is a diagonal matrix , @xmath9 $ ] . note that hyperbolic metamaterials are a special kind of uniaxial medium satisfying @xmath10 . as mentioned before , we should employ the second kind of fdt because this is a non - equilibrium problem . to solve dgf in planar structures , it is convenient to work in the wavevector space . dgf in vacuum@xcite is ( @xmath11 ) @xmath12 here we define @xmath13 is the normalized wave - vector of upward waves ( @xmath14 ) in free space , @xmath15 , @xmath16 , @xmath17 , and @xmath18 . @xmath19 is the unit direction vector of s - polarized waves , @xmath20 is the unit direction vector of p - polarized waves . correspondingly and for later use , @xmath21 is the normalized wave - vector of downward waves ( when @xmath22 ) , @xmath23 same with @xmath24 , and @xmath25 . the dgf relating thermally generated random currents inside the medium in the lower space to the fields in upper space vacuum is @xmath26 here , @xmath27 , @xmath28 . @xmath29 , and @xmath30 which are the unit direction vectors of s- and p - polarized waves inside the unaxial medium , respectively . note the transmission coefficients incident from the vacuum side should be in terms of the electric fields , @xmath31 to calculate the magnetic fields , we should evaluate @xmath32 , which can be easily done in the wavevector space . the curl operator will work on the first vector of @xmath33 , @xmath34 the free space energy density is defined by @xmath35 where the prefactor 2 accounts for the negative frequency counterpart . following the formalism in ref . , we define @xmath36 one can then find @xmath37 inserting the expressions of @xmath38 and @xmath39 , we have @xmath40 the integration on @xmath41 can be easily done . further by taking the imaginary part of the dispersion relation @xmath42 for s- and p - polarized waves , this result can be simplified as @xmath43 here @xmath44 is the energy density of blackbody.@xmath45 and @xmath46 are the standard reflection coefficients given by @xmath47 the propagating wave part @xmath48 in eq . ( [ general expression ] ) is the far field emissivity , equivalent to kirchhoff s law . correspondingly , the evanescent wave part can be interpreted as kirchhoff s law in the near field and @xmath49 is the near field emissivity@xcite , which is widely used in heat transfer problems . @xmath49 is also proportional to the near field local density of states ( ldos ) proposed in ref . and is related to the tunneling and subsequent absorption of energy carried by evanescent waves . recently extensive theoretical and experimental works have demonstrated the ability of hmms to enhance the near field ldos@xcite . thus we expect the use of hmms in thermal and energy management . we can use the above definitions to compare the energy density in the near - field of the hyperbolic media to any other control sample . a pertinent question is about how much energy density is in matter degrees of freedom as opposed to the fields . this is difficult to answer inside the medium but can be done unambiguously in the near - field . in the high - k approximation , where the wavevector parallel to the interface @xmath50 is sufficiently large , the near - field energy density is governed by the tunneling parameter which we define as the imaginary part of the p - polarized reflection coefficient . thus studying the behavior of this tunneling parameter sheds light on the near - field energy density . in the low loss limit , the reflection for p - polarized waves incident on an interface between vacuum and hmm can be expressed by@xcite @xmath51 while for an isotropic medium , the high - k approximation gives @xmath52 the most striking difference between the above equations is that for a conventional isotropic medium the near - field energy density is completely dominated by the imaginary part of the dielectric constant . these fluctuations disappear in the low loss limit and can be attributed to matter degrees of freedom . this is because the imaginary part of the dielectric constant which governs field fluctuations also characterizes the irreversible conversion of electromagnetic energy into thermal energy of matter degrees of freedom . on the other hand , the hyperbolic medium shows near - field fluctuations arising from high - k modes completely indpendent of material losses and the energy resides in the field . let us analyze what would happen at mid - infrared frequencies where phonon polaritonic materials can give rise to this low loss high - k limit for hyperbolic media . we clearly see from eq . [ high_k_iso ] that the near field emissivity would be very small when the frequency is away from the surface phonon polariton resonance ( sphpr ) frequency where @xmath53 . however , for hmms made of phonon polaritonic materials and dielectrics , the near field emissivity ( eq . [ high_k_hmm ] ) can be comparably large in broad frequency region , though in this approximation its magnitude can not exceed one . note here we do not account for surface wave resonances which can change the picture considerably especially if one wants to optimize near - field heat transfer@xcite . our aim is to focus on the bulk modes only . in this section we will consider multilayered structures . in the field of metamaterials , multilayered structures are widely used to achieve effective uniaxial media . the aim here is to go beyond effective medium theory and calculate the exact thermal emission from multilayered structures using the second kind of fdt . we assume that the medium in all layers is isotropic and non - magneto - optical for simplicity . to find dgfs relating the random currents in each layer to the vacuum region , we will follow the method in ref . . firstly assuming the current source is in the vacuum region , we can calculate the fields induced by the source in all the layers by transfer matrix method which matches the boundary conditions at all the interfaces . thus the dgfs with source in the vacuum region are ready to be employed . next we use the reciprocal property of the dgf to achieve dgf when the sources are in the lower space . dgf in the vacuum region ( @xmath54 ) is @xmath55 dgf in the intermediate slabs are @xmath56 dgf in the last layer is @xmath57 note in the last layer we only have the downward waves , namely , the transmission . the boundary conditions give@xcite @xmath58 for s - polarized waves , and @xmath59 for p - polarized waves . following the same steps as in the uniaxial case , the final expression is @xmath60 where @xmath61 is the total number of layers in the structure . to simplify the above result , we first note that the integral @xmath62\big \|_{z_{l}}^{z_{l-1 } } \nonumber \\ & = q_l(z_{l-1})-q_l(z_{l } ) , \end{aligned}\ ] ] which is valid for all layers . from the boundary condition , we have @xmath63 thus we find @xmath64 for the last term , @xmath65 , so @xmath66 , and in our convention , @xmath67 . the final result is @xmath68= \nonumber \\ \left\ { \begin{gathered } ( 1-\|r^s\|^2)\|k_{z0}\|,\qquad k_\rho < k_0 \\ 2{\operatorname{im}}(r^s)\|k_{z0}\|,\qquad k_\rho > k_0 \end{gathered } \right . . \end{aligned}\ ] ] this is the contribution from s - polarized waves . for p - polarized waves , the corresponding identity is @xmath69\big \|_{z_{l}}^{z_{l-1}}\end{aligned}\ ] ] then the contribution from p - polarized waves can be evaluated in the similar way . the final expression for thermal emission from a half space multilayered structure will be given by eq . [ general expression ] . the reflection coefficients should be that of the whole structure . if we are interested in a slab inside vacuum rather than a half space structure , we can eliminate the contribution from the last layer vacuum part . to do so , in eq . ( [ last layer ] ) , for the last layer @xmath70 and @xmath71 , the right hand side is therefore @xmath72 , which vanishes for evanescent waves . subtracting this term from eq . [ general expression ] gives the thermal emission from a multilayered slab inside vacuum , @xmath73 the above expression can be also obtained by replacing @xmath48 in eq . [ general expression ] with @xmath74 , which is consistent with kirchoff s law . we now describe another approach to evaluating the near - field energy density near metamaterials using the scattering matrix approach . however , first we will discuss a few important points related to the concept of the thermal environment . we note that when the lower space is vacuum , the reflection coefficients are zero . as a result of eq . ( [ general expression ] ) , the contribution from the evanescent waves part is zero while that from the propagating waves is nonzero . however , this is not very intuitive from fdt . the reason is that losses of vacuum i.e. @xmath75 of vacuum is zero and from the second kind of fdt , the correlation function of random currents of vacuum should be zero , suggesting a zero field correlation . it turns out that for an unbounded vacuum region , we should add an infinitesimal imaginary part to @xmath76 , integrate over the region and then take the limit of the imaginary part to be zero in the final expression@xcite . this is needed to preserve causality requirements . in the derivation of eq . ( [ general expression ] ) , we have integrated the source region @xmath41 from @xmath77 to @xmath78 . however , for a vacuum gap with any finite width , the final fields correlation originating from the gap can be shown to be zero@xcite . for this reason , fluctuations in vacuum can be interpreted to come from infinity . it is then natural to think about the thermal emission from the upper space vacuum region as well . if the vacuum region is also at temperature t , the system is at global thermal equilibrium . therefore we can employ the first kind of fdt to calculate the thermal energy density . this approach is used in ref . to define the local density of states . here we directly cite the final result , @xmath79 ) \nonumber \\ & + \left . { \int_{{k_0}}^\infty { \frac{{k_\rho ^3d{k_\rho } } } { { k_0 ^ 3\left\| { { k_{z0 } } } \right\|}}{e^ { - 2\operatorname{im } ( { k_{z0}})z}}(\operatorname{im } ( { r^s } ) + \operatorname{im } ( { r^p } ) ) } } \right\}\end{aligned}\ ] ] note again that the contribution from evanescent waves equals that of eq . ( [ general expression ] ) , implying no evanescent waves contribution from the upper space vacuum region . however , in non - equilibrium , to determine electromagnetic fields induced by every random current inside the medium using second kind of fdt is quite laborious . we note from the second kind of fdt that the currents are not spatially correlated , which suggests that the thermal emission from different spatial regions can be calculated separately . in thermal equilibrium , we can calculate the thermal energy density by the first kind of fdt . thus if we can calculate the thermal emission from the upper space vacuum part at temperature t , thermal emission only from the lower space can be achieved by excluding the vacuum part from the total thermal energy density . the electric field generated by the upper half vacuum space can be written as@xcite @xmath80 where @xmath81 @xmath82 and @xmath83 are the field amplitude for s and p - polarized waves , respectively . the operator @xmath84 satisfies the correlation function@xcite , @xmath85 the coefficient @xmath86 can be read directly from fdt and the free space dgf , @xmath87 which vanishes for evanescent waves . these fluctuations from the upper vacuum region shines on the interface and get reflected . the total fields due to fluctuations in the vacuum part are @xmath88 the magnetic fields can be calculated using eq . ( [ eq : vacuum fluctuation ] ) and maxwell equations . then one can find the energy density due to the fluctuations in the upper space vacuum , @xmath89\big \}\end{aligned}\ ] ] subtracting eq . ( [ thermal vacuum ] ) from eq . ( [ total emission ] ) , we recover the expression by the second kind of fdt . from the definition of the cross - spectral density tensor @xmath90 one can find the spatial coherence due to fluctuations in the upper space vacuum , @xmath91\end{aligned}\ ] ] where @xmath92 , @xmath93 and @xmath94 is used ; @xmath95 is the zeroth order bessel function of the first kind . further , from eq . ( [ first kind ] ) , the first kind of fdt , we have @xmath96 + \int_{k_0}^{\infty } \frac{{k_\rho^3 d{k_\rho } } } { k_0 ^ 3 \|k_{z0}\| } \nonumber \\ & j_0(k_\rho d ) { \operatorname{im}}(r^p)e^{-2{\operatorname{im}}(k_{z0})z } \big \}\end{aligned}\ ] ] then the contribution from the lower space structure is @xmath97 only p - polarized waves contributes to @xmath98 since s - polarized waves do not have @xmath99 components . once again , if the structure is a multilayered slab in vacuum , the contribution from the lower vacuum space can be evaluated using the scattering matrix method in a similar way to the upper vacuum space . the fields due to the vacuum fluctuations in the lower space transmit through the planar structure , @xmath100 it is clear that the contributing energy density will be proportional to the @xmath101 , so that we recover the result of eq . ( [ thermal slab ] ) . note that due to the reciprocal property , the transmission coefficients from two sides of the structure should be identical . generally speaking , considering a single object in thermal equilibrium , the energy density can be determined by the first kind of fdt , which is simply a single scattering event . to find the contribution from the object only , we can exclude the contribution from the environment , which can be also expressed by the scattering matrix of the object . if there are several objects at different temperatures , we can first decide the thermal emission from one specific object in the absence of other objects and then build the scattering part from other objects , in which procedure the temperatures of the other objects and the environment are assumed to be zero . note this is the basic idea of m.kardar and co - authors in sequent works@xcite . beyond the multilayered structures considered here , the authors also give the scattering matrix of various geometries including sphere and cylinder . for more complicated objects , numerical methods are also well developed.@xcite ( a ) schematic of the multilayered structure and the coordinates . the spatial coherence are calculated between @xmath102 and @xmath103 . ( b ) effective permittivities of a sio@xmath104-sic multilayered structure , where the fill fraction of sic is 0.4 . only real part of the permittivity is plotted . the insets from left to right , denote the iso - frequency dispersion of dielectric , type ii hmm and type i hmm . ] normalized far field thermal emission of a 3@xmath105 m sio@xmath104-sic multilayered structure , with fill fraction of sic is 0.4 . ] there are multiple approaches to achieving hyperbolic dispersion@xcite . two of the prominent geometries consists of 1d or 2d periodic metal - dielectric structures . we consider here a multilayer combination of silicon dioxide ( sio@xmath104 ) and silicon carbide ( sic ) which has a metallic response in the reststrahlen band due to phonon polaritons ( @xmath106 between @xmath107 hz and @xmath108 hz , the transverse and longitudinal optical phonon resonance frequencies ) . the permittivity of sic is given by @xmath109 , where @xmath110 is the frequency of operation , @xmath111 and @xmath112 hz . we note that this realization formed the testbed for the first complete characterization of the modes of hyperbolic media due to their low loss as compared to plasmonic media@xcite . the modes of this hmm can be excited at relatively lower temperatures ( 400 - 500k ) when the peak of black body emission lies within the reststrahlen band of sic . wavevector resolved thermal emission ( normalized to blackbody emission into the upper space ) from a sio@xmath104-sic multilayered structure calculated by ( a ) transfer matrix method and ( b ) emt at z=200 nm . the thermal emission is normalized to the black body emission to the upper half - space and in log scale . the structure consists of 40 layers of sio@xmath104/sic , 30nm/20 nm achieving a net thickness of 1@xmath105 m . the presence of high - k modes are clearly evident in both the emt calculation and the multilayer practical realization which takes into account all non - idealities due to dispersion , losses , finite unit cell size and finite sample size . the bright curves denote the enhanced thermal emission due to high - k modes in the hmm . in the practical multilayered structure , the high - k modes come from the coupled short range surface phonon polaritons at the silicon carbide and silicon dioxide interfaces . ] to understand the thermal properties of phonon - polaritonic hyperbolic metamaterials we need to focus only on the reststrahlen band of sic where it is metallic . the multilayer structure ( see schematic in fig . [ fig : emt_para](a ) ) shows a host of different electromagnetic responses as predicted by effective medium theory @xmath113 and @xmath114 , here @xmath115 is the fill fraction of the metallic medium@xcite . we classify the effective uniaxial medium@xcite using the isofrequency surface of extraordinary waves which follow @xmath116 and the media are hyperboloidal only when @xmath117 . we can effectively achieve a type i hyperbolic metamaterial with only one negative component in the dielectric tensor ( @xmath118 , @xmath119 ) , type ii hyperbolic metamaterial with two negative components ( @xmath120 , @xmath121 ) , effective anisotropic dielectric ( @xmath118 , @xmath121 ) or effective anisotropic metal ( @xmath120 , @xmath119 ) . in fig . [ fig : emt_para](b ) , we plot the effective permittivities of a sio@xmath104-sic multilayered structure with the fill fraction 0.4 and label the two hyperbolic regions . as the purpose of this work is to examine how extraordinary waves in hmms impact thermal emission properties , we only consider p - polarized waves in our numerical simulations . we first characterize the thermal emission of a hmm slab in the far field . this is extremely important for experiments currently being pursued in multiple groups . we clearly observe two peaks in fig . [ fig : far field ] in agreement with the previous work on epsilon - near - zero and epsilon - near - pole resonances for thermal emission@xcite . the right one occurs when @xmath122 is close to zero . from the displacement field boundary condition , @xmath123 , when @xmath124 , the fields inside hmm @xmath125 should be very large . thus large absorption is expected at this epsilon near zero region . the epsilon - near - pole resonance results in narrowband thermal emission due to the increase in the imaginary part of the dielectric constant in this enp spectral region . the most critical aspect is the direction of the dielectric tensor components which show enz or enp@xcite . an enz in the component parallel to the interface or an enp perpendicular to the interface does not show such effects . wavevector resolved thermal emission ( normalized to blackbody emission into the upper space ) from ( a ) a 3@xmath105 m thickness hmm slab and ( b ) a 30@xmath105 m thickness hmm slab . the fill fraction of sic is 0.4 , same as the 1@xmath105 m hmm slab . the two hyperbolic regions where the thermal emission is enhanced are evident . the modes supported by 3@xmath105 m thickness slab are denser than that of 1@xmath105 m slab and the modes supported by the 30@xmath105 m slab are almost continuous . ] ( a ) normalized thermal emission from slabs with various thicknesses . the dashed black line is calculated using transfer matrix method while the solid lines are calculated using emt parameters , where dm in the legend means the top layer of sio@xmath104(dielectric)-sic(metal ) multilayers is sio@xmath104 . despite the clear difference of the density of modes supported by the slabs shown in fig . [ fig:1_emt_dm ] and [ fig:3_30_emt ] , the thermal emission spectrum are interestingly in good agreement . the two main peaks where the thermal emission are largely enhanced are due to the high - k states in the two hyperbolic regions . ( b ) wavevector resolved thermal emission at @xmath126hz . the sharp peaks on the left ( @xmath127 ) are the surface modes . when @xmath128 , the curve for 30@xmath105 m slab is almost flat with no oscillations , while that of 1@xmath105 m and 3@xmath105 m slabs show the discrete modes denoted by crests and troughs . ] thermal emission by a 30@xmath105 m sic slab . the red bright curve represents the dispersion of the sphp mode between the vacuum and sic interface since the slab is very thick . ] spatial coherence of ( a ) a 30@xmath105 m sic slab and ( b ) a 30@xmath105 m hmm slab at 0.2@xmath105 m and 1@xmath105 m from the surface with @xmath126hz and @xmath129hz . ( a ) at @xmath126hz , the sic slab supports a single degenerate sphp mode . as a result , sic slab has large spatial coherence at both 0.2@xmath105 m and 1@xmath105 m . at @xmath129hz , the sphp resonance frequency where @xmath130 , this frequency corresponds to a bright horizontal line in the sphp dispersion curve shown in fig . [ fig : sic_phase ] . this means at this frequency , multi - modes with different wavevectors can be thermally excited . thus the spatial coherence is poor both at 0.2@xmath105 m and 1@xmath105 m . ( b)at @xmath126hz , the hmm slab supports high - k states besides the sphp mode . at 0.2@xmath105 m , the high - k states contribute a lot to the fluctuating electric fields , and consequently the spatial coherence is poor . but when the distance becomes larger at 1@xmath105 m , the high - k states will not reach that far because of their large wavevector @xmath50 . thus the electric fields will be dominated by the surface mode which has smaller @xmath50 . the spatial coherence length is large due to this dominant surface mode . at @xmath129hz , the hmm slab can only supports multiple high - k states , and unlike the type ii hmm region , there is no lower bound for the high - k wavevectors . thus the spatial coherence is poor both at 0.2@xmath105 m and 1@xmath105 m . ] here we analyze the near - field thermal emission from multilayer hyperbolic media@xcite . we first focus on how thermal emission will depend on the thickness of the slabs . in fig . [ fig:1_emt_dm ] , we plot the wavevector resolved thermal emission from a structure consists of 40 layers of sio@xmath104/sic , 30nm/20 nm achieving a net thickness of 1@xmath105 m . we clearly see multiple discrete high - k modes in both the type i and type ii hyperbolic region . note the thickness 1@xmath105 m is about one tenth of the operating wavelength , so these high - k modes will not occur in conventional isotropic dielectrics . the excellent agreement between the emt prediction and the practical multilayered structure is seen , which validates the use of emt in our structure . further , we increase the thickness of the slab to 3@xmath105 m and 30@xmath105 m while keeping the same unit cell . the waveguide modes will be denser as expected . at the thickness of 30@xmath105 m , the high - k modes are almost continuous and result in two bright bands in fig . [ fig:3_30_emt](b ) . this is close to the bulk metamaterial limit . we show the thermal emission spectrum in fig . [ fig : emission](a ) for various thicknesses of the metamaterial . the two main peaks are due to the high - k modes in the hyperbolic region . in fig . [ fig : emission](b ) , we plot the wavevector resolved thermal emission at a specific frequency @xmath126hz within the type ii hyperbolic region where the structure supports both surface mode and high - k modes . the sharp peaks at the left are due to the surface mode while the high - k modes emerge at larger @xmath50 . in the high - k modes region , the curve for 30@xmath105 m slab is almost flat indicative of a continuum of high - k modes . in contrast , the curves of 1@xmath105 m and 3@xmath105 m slabs clearly show the existence of discrete high - k waveguide modes featured by crests and troughs . surface waves can lead to large spatial coherence length in the near field@xcite . to see this , we first show in fig . [ fig : sic_phase ] the wavevector resolved thermal emission from a 30@xmath105 m thick sic slab . the bright curve gives the dispersion of surface phonon polariton ( sphp ) between the vacuum and sic interface . note we will not see the splitting of the vacuum - sic interface sphp mode into long range and short range modes since 30@xmath105 m is in the order of several operating wavelengths . in the time domain , the temporal coherence is best for monochromatic waves . thus for the spatial coherence , one can imagine it will be favorable if a single wavevector dominates the fields among all the wavevectors . this is indeed the case for surface waves . in fig . [ fig : spatial](a ) , we plot the spatial coherence of the sic slab at @xmath131hz and @xmath132hz . at the frequency @xmath131hz , the sphp mode wavevector @xmath50 is about @xmath133 . large spatial coherence length is seen at both 0.2@xmath105 m and 1@xmath105 m from the interface . however , near the surface phonon polariton resonance ( sphpr ) frequency @xmath132hz where @xmath134 , the mode dispersion curve is almost a horizontal line , which means that multiple modes with different wavevectors can be thermally excited . thus a poor spatial coherence is expected . in fig . [ fig : spatial](a ) , the spatial coherence is poor at at both 0.2@xmath105 m and 1@xmath105 m from the interface . this feature could be used to determine the resonance frequency . hyperbolic metamaterials can support multiple high - k modes . therefore the spatial coherence length should not be long in the hyperbolic region . this is true for type i hmm . in fig . [ fig : spatial](b ) , we plot @xmath98 at @xmath132hz , where the multilayered structure effectively behaves in the type i hyperbolic region . the spatial coherence lengths are only a fraction of the operating wavelength at both 0.2@xmath105 m and 1@xmath105 m from the interface . but the situation for type ii hyperbolic region is interestingly different . for a hmm slab in the type ii hyperbolic region ( @xmath120 , @xmath121 ) , the slab can support a surface wave mode as well as multiple high - k modes . thus we have two sets of modes that can result in a unique interplay of spatial coherence effects . furthermore , these modes are separated in wavevector space because of the lower bound of the high - k states in type ii hyperbolic region@xcite . high - k modes are confined to the surface better than suface waves and these high - k waves will dominate at a shorter distance from the interface . we choose @xmath131hz within the type ii hyperbolic region to confirm this point . at distance 0.2@xmath105 m , the spatial coherence is very poor . however , at a larger distance 1@xmath105 m , the fluctuating fields have large spatial coherence length . this is because at this distance , the contribution from surface wave mode dominates the electric fields while the high - k states rarely contribute to the fields . this distance dependence behavior can have applications such as obtaining the modes distribution at a given frequency . until now , we have fixed the fill fraction to be 0.4 . it is useful to examine the structure s behavior at various fill fractions . in fig . [ fig : ttt](a ) , we plot the optical phase diagram@xcite of this metamaterial which shows the isofrequency surfaces achieved at different frequencies and fill fractions of sic . the phase diagram is classified as effective dielectric , effective metal , type i and type ii hmm as introduced before@xcite . [ fig : ttt](b ) shows the thermal energy density ( normalized to black body radiation into the upper half space ) evaluated using rytov s fluctuational electrodynamics for an effective medium slab at a distance of z=200 nm from the metamaterial . it is seen that the regions of hyperbolic behavior exhibit super - planckian thermal emission in agreement with our previous analytical approximation , but here we will go beyond effective medium theory and consider practical structures . the role of the surface waves is very important and can lead to significant deviations when the unit cell size is not significantly subwavelength@xcite . the macroscopic homogenization utilized to define a bulk electromagnetic response is valid when the wavelength of operation exceeds the unit cell size ( @xmath135 ) . however , even at such wavelengths if one considers incident evanescent waves on the metamaterial the unit cell microstructure causes significant deviations from emt . this is an important issue to be considered for quantum and thermal applications where the near - field properties essentially arise from evanescent wave engineering ( high - k modes)@xcite . for the multilayer hmm , at distances below the unit cell size , the thermal emission is dominated by evanescent waves with lateral wavevectors @xmath136 . since this is above the unit - cell cut off of the metamaterial , the high - k modes do not contribute to thermal emission at such distances . it is therefore necessary to consider thermal emission from a practical multi - layer structure taking into account the layer thicknesses . this is shown in fig . [ fig : ttt](c ) and fig . [ fig : ttt](d ) . the unit cell size is 200 nm , and we consider a semi - infinite multilayer medium using the formalism outlined in ref . . an excellent agreement is seen of the optical phases of the multilayer structure with the emt calculation . this work shows that extension of equilibrium and non - equilibrium fluctuational electrodynamics to the case of metamaterials can lead to novel phenomena and applications in thermal photonics . we presented a unified picture of far - field and near - field spectra for experimentalists and also introduced the near - field spatial coherence properties of hyperbolic media . we have analyzed in detail thermal topological transtions and super - planckian thermal emission in practical phonon - polaritonic hyperbolic metamaterials . we paid particular attention not only to the effective medium approximation but discussed all non - idealities limiting the super - planckian thermal emission from hmms . we have provided practical designs to experimentally measure and isolate our predicted effect . our work should lead to a class of thermal engineering applications of metamaterials . z. jacob wishes to acknowledge discussions with e. e. narimanov . this work was partially supported by funding from helmholtz alberta initiative , alberta innovates technology futures and national science and engineering research council of canada . 89ifxundefined [ 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 @noop _ ( , ) @noop * * , ( ) link:\doibase 10.1103/physrevb.4.3303 [ * * , ( ) ] @noop _ _ ( , ) @noop _ _ ( , ) link:\doibase 10.1364/oe.19.0a1088 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.85.1548 [ * * , ( ) ] \doibase doi:10.1063/1.4754616 [ * * , ( ) ] link:\doibase 10.1016/j.surfrep.2004.12.002 [ * * , ( ) ] \doibase doi:10.1063/1.2234560 [ * * , ( ) ] link:\doibase 10.1103/physrevb.86.085432 [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.195459 [ * * , ( ) ] link:\doibase 10.1080/10893950290053321 [ * * , ( ) ] link:\doibase 10.1103/physrevb.70.125101 [ * * , ( ) ] link:\doibase 10.1038/nphoton.2009.144 [ * * , ( ) ] link:\doibase 10.1103/physrevb.63.205404 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.79.1291 [ * * , ( ) ] link:\doibase 10.1088/0953 - 8984/11/35/301 [ * * , ( ) ] link:\doibase 10.1103/physrevb.81.165119 [ * * , ( ) ] link:\doibase 10.3367/ufne.0179.200905a.0449 [ * * , ( ) ] link:\doibase 10.1016/j.ijheatmasstransfer.2005.09.037 [ * * , ( ) ] link:\doibase 10.1063/1.2963195 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.104.154301 [ * * , ( ) ] \doibase doi:10.1063/1.2905286 [ * * , ( ) ] link:\doibase 10.1021/nl901208v [ * * , ( ) ] link:\doibase 10.1103/physrevlett.107.045901 [ * * , ( ) ] link:\doibase 10.1021/nl301708e [ * * , ( ) ] link:\doibase 10.1038/nature05265 [ * * , ( ) ] \doibase doi:10.1063/1.1575936 [ * * , ( ) ] link:\doibase 10.1021/nl204201 g [ * * , ( ) ] link:\doibase 10.1103/physreva.80.042102 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.106.210404 [ * * , ( ) ] link:\doibase 10.1103/physrevb.86.115423 [ * * , ( ) ] link:\doibase 10.1103/physrevd.87.025016 [ * * , ( ) ] link:\doibase 10.1209/0295 - 5075/95/61002 [ * * , ( ) ] link:\doibase 10.1103/physreva.84.042102 [ * * , ( ) ] link:\doibase 10.1103/physrevd.80.085021 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.97.223203 [ * * , ( ) ] link:\doibase 10.1103/physreva.76.013810 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.96.160401 [ * * , ( ) ] link:\doibase 10.1109/jproc.2012.2191749 [ * * , ( ) ] link:\doibase 10.1103/physrevb.86.220302 [ * * , ( ) ] link:\doibase 10.1103/physrevb.88.054305 [ * * , ( ) ] @noop _ _ ( , ) link:\doibase 10.1038/nphoton.2006.49 [ * * , ( ) ] link:\doibase 10.1126/science.1096796 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.85.3966 [ * * , ( ) ] link:\doibase 10.1070/pu1968v010n04abeh003699 [ * * , ( ) ] link:\doibase 10.1103/physrevb.79.035407 [ * * , ( ) ] link:\doibase 10.1088/1464 - 4258/11/11/114003 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.102.023901 [ * * , ( ) ] link:\doibase 10.1038/nphoton.2007.28 [ * * , ( ) ] link:\doibase 10.1126/science.1125907 [ * * , ( ) ] link:\doibase 10.1126/science.1133628 [ * * , ( ) ] link:\doibase 10.1126/science.1108759 [ * * , ( ) ] link:\doibase 10.1364/oe.14.008247 [ * * , ( ) ] link:\doibase 10.1088/2040 - 8978/14/6/063001 [ * * , ( ) ] link:\doibase 10.1155/2012/452502 [ * * ( ) , 10.1155/2012/452502 ] link:\doibase 10.1126/science.1219171 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.90.077405 [ * * , ( ) ] link:\doibase 10.1364/josab.21.001032 [ * * , ( ) ] link:\doibase 10.1103/physrevb.71.201101 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.109.104301 [ * * , ( ) ] link:\doibase 10.1364/oe.21.015014 [ * * , ( ) ] link:\doibase 10.1364/oe.21.014988 [ * * , ( ) ] link:\doibase 10.1063/1.4800233 [ * * , ( ) ] link:\doibase 10.1016/0030 - 4018(82)90402 - 3 [ * * , ( ) ] link:\doibase 10.1103/physreva.29.1991 [ * * , ( ) ] link:\doibase 10.1103/physreva.85.053842 [ * * , ( ) ] link:\doibase 10.1364/oe.18.022734 [ * * , ( ) ] link:\doibase 10.1364/oe.21.000a96 [ * * , ( ) ] link:\doibase 10.1088/2040 - 8978/15/10/105101 [ * * , ( ) ] link:\doibase 10.1364/oe.20.008100 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.82.1660 [ * * , ( ) ] link:\doibase 10.1016/s0030 - 4018(00)01048 - 8 [ * * , ( ) ] link:\doibase 10.1007/s00340 - 008 - 3204 - 2 [ * * , ( ) ] link:\doibase 10.1103/physreve.76.016601 [ * * , ( ) ] http://www.degruyter.com/view/j/apsrt.2008.58.issue-5/v10155-010-0092-x/v10155-010-0092-x.xml [ * * , ( ) ] @noop _ _ , vol . ( , ) @noop _ _ , vol . ( , ) @noop _ _ ( , ) link:\doibase 10.1103/physrevlett.105.234301 [ * * , ( ) ] link:\doibase 10.1007/s00340 - 010 - 4096 - 5 [ * * , ( ) ] http://arxiv.org/abs/1311.5812 [ ( ) ] http://books.google.ca/books?id=devtkqeacaaj[__ ] , landau and lifshitz course of theoretical physics ( , ) link:\doibase 10.1103/physrevb.68.245405 [ * * , ( ) ] link:\doibase 10.1103/physreve.85.046603 [ * * , ( ) ] link:\doibase 10.1103/physrevb.85.165104 [ * * , ( ) ] link:\doibase 10.1016/j.jqsrt.2013.04.017 [ * * , ( ) ]	we give a detailed account of equilibrium and non - equilibrium fluctuational electrodynamics of hyperbolic metamaterials . we show the unifying aspects of two different approaches ; one utilizes the second kind of fluctuation dissipation theorem and the other makes use of the scattering method . we analyze the near - field of hyperbolic media at finite temperatures and show that the lack of spatial coherence can be attributed to the multi - modal nature of super - planckian thermal emission . we also adopt the analysis to phonon - polaritonic super - lattice metamaterials and describe the regimes suitable for experimental verification of our predicted effects . the results reveal that far - field thermal emission spectra are dominated by epsilon - near - zero and epsilon - near - pole responses as expected from kirchoff s laws . our work should aid both theorists and experimentalists to study complex media and engineer equilibrium and non - equilibrium fluctuations for applications in thermal photonics .
fix an algebraically closed field @xmath35 . let @xmath12 be a nonsingular projective curve defined over @xmath35 of genus @xmath36 . hence , when @xmath37 , the curve @xmath12 is a compact riemann surface of genus @xmath38 . let @xmath13 be an invertible sheaf on @xmath12 of degree @xmath39 , and let @xmath40 be the line bundle associated to @xmath13 ( as in ( * ? ? ? * ii , ex.5.18 ) ) , where @xmath41 is the dual of @xmath13 . this is the so - called total space of @xmath13 . a section of @xmath42 is a morphism @xmath43 such that the composition map @xmath44 is the identity . the space of sections can be identified with @xmath45 . since it is better to deal with a projective surface , we naturally compactify all fibers , so that we work with a @xmath46-bundle . let @xmath47 be the @xmath46-bundle associated to @xmath48 over @xmath12 , as in ( * ? ? ? * ii , ex.7.10 ) . the nonsingular projective surface @xmath49 contains @xmath50 as an open set , such that the curve @xmath51 is a section of @xmath52 with self - intersection @xmath53 . it is easy to see that @xmath54 is the only irreducible curve with negative self - intersection in @xmath55 . this surface is a particular case of a geometrically ruled surface over @xmath12 ( * ? ? ? * v , section 2 ) , and it is in its normalized form ( * ? ? ? * v , proposition 2.8 ) . we denote by @xmath56 the fiber over a point @xmath57 , or just @xmath58 when we consider its numerical class . any element in @xmath59 can be written as @xmath60 with @xmath61 , and @xmath62 . any element in @xmath63 can be written as @xmath64 with @xmath65 @xcite . let @xmath15 , and let @xmath66 . consider the invertible sheaf @xmath67 on @xmath46 . then , the surface @xmath68 is the hirzebruch surface @xmath69 . when @xmath70 , @xmath71 corresponds to the blow - up at a point of @xmath0 ( * ? ? ? * v , exa.2.11.4 ) , and @xmath54 is the @xmath72-curve . [ e11 ] the main objects are the following . let @xmath73 be an integer . let @xmath12 be a nonsingular projective curve , and let @xmath13 be an invertible sheaf on @xmath12 of degree @xmath66 . an * arrangement of @xmath1 sections * is a labeled set of @xmath1 distinct sections @xmath74 of @xmath75 such that @xmath76 for all @xmath77 , and @xmath78 . from now on , we denote @xmath79 by @xmath80 . [ d11 ] in particular , @xmath81 , and @xmath82 for all @xmath83 , and so these arrangements are indeed formed by sections of @xmath84 . the condition @xmath78 implies that @xmath13 is base - point free . to see this , take a point @xmath85 , and consider the corresponding fiber @xmath56 . since @xmath78 , there are two sections @xmath86 , @xmath87 such that @xmath88 . let @xmath89 be the map defining the section @xmath87 . then , @xmath90 , and @xmath91 is given by an effective divisor on @xmath12 not supported at @xmath92 . this tell us that @xmath93 with @xmath94 base - point free effective divisor . if @xmath95 is an arrangement as in definition [ d11 ] , but with @xmath96 , then we can apply elementary transformations ( see ( * ? ? ? * v , exa.5.7.1 ) ) at each of the points in @xmath97 to obtain a new arrangement of @xmath1 sections in @xmath98 for some @xmath99 . after repeating this process a finite number of times , we arrive to an arrangement @xmath100 in @xmath98 with @xmath101 . if @xmath102 , then @xmath103 since , as we showed above , @xmath104 for some effective divisor @xmath94 . in this case , @xmath105 , and the arrangement is trivially formed by a collection of @xmath1 horizontal " fibers ( it just corresponds to an arrangement of @xmath1 points in @xmath46 ) . if @xmath106 is a collection of arbitrary @xmath1 sections in @xmath80 , we perform elementary transformations on the points in @xmath107 for all @xmath83 , and we repeat this process until all sections are disjoint from the new curve @xmath108 in @xmath98 ( proper transform of @xmath54 ) . in this way , * arbitrary arrangements of sections * can always be considered , after some elementary transformations , as the ones in definition [ d11 ] . we now define the morphisms between our objects . fix an integer @xmath73 . let @xmath109 be nonsingular projective curves , and let @xmath110 be invertible sheaves of positive degrees on @xmath109 respectively . let @xmath111 be arrangements of @xmath1 sections in @xmath112 respectively . a * morphism of arrangements * is the existence of a finite map @xmath113 , and a commutative diagram @xmath114_{\pi } \ar[r]^{g } & \p_{c'}(\l ' ) \ar[d]_{\pi ' } \\ c \ar[r]^{g } & c ' } $ ] so that @xmath80 is isomorphic to the base change by @xmath38 , and @xmath115 for all @xmath83 . if @xmath38 is an isomorphism , then the arrangements are said to be isomorphic . [ d12 ] in particular , a curve @xmath12 with an automorphism @xmath38 produces isomorphic arrangements via the pull back of @xmath38 . a morphism of arrangements satisfies @xmath116 and @xmath117 . we have @xmath118 and @xmath119 for all @xmath83 . [ l11 ] since @xmath120 , we have that @xmath121 . we know that @xmath122 and @xmath123 ( see ( * ? ? ? * ii , proposition 7.11 ) ) . by flat base change ( * iii , proposition 9.3 ) , we have @xmath124 and so @xmath125 . therefore , @xmath126 , and so @xmath127 , and @xmath128 for all @xmath83 . one wants to consider arrangements of sections which do not come from others via base change , and so the following definition . let us fix the data @xmath129 as above . an arrangement of @xmath1 sections @xmath5 is said to be * primitive * if whenever we have an arrangement @xmath100 for some data @xmath130 , and a morphism @xmath38 as in definition [ d12 ] , then @xmath38 is an isomorphism . the set of * isomorphism classes * of primitive arrangements is denoted by @xmath131 . this is clearly independent of the isomorphism class of @xmath12 and @xmath13 . [ d13 ] for instance , if @xmath13 has a base - point , then @xmath132 . let @xmath73 be an integer . an * arrangement of @xmath1 lines * in the plane is a set of @xmath1 labeled lines @xmath133 in @xmath0 such that @xmath134 . as in @xcite , we introduced ordered pairs @xmath135 , where @xmath5 is an arrangement of @xmath1 lines , and @xmath136 is a point in @xmath137 . if @xmath135 and @xmath138 are two such pairs , we say that they are isomorphic if there exists an automorphism @xmath139 of @xmath0 such that @xmath140 for every @xmath83 , and @xmath141 . given @xmath135 , we blow up the point @xmath136 to obtain an arrangement of @xmath1 sections for the data @xmath142 , and given such an arrangement of sections , we blow down @xmath54 to get a pair @xmath135 , where @xmath136 is the image of @xmath54 and @xmath5 is formed by the images of the sections . one sees that the set of pairs up to isomorphism of pairs is precisely @xmath143 . by lemma [ l11 ] , any arrangement of @xmath144 is primitive ( degree considerations ) . this is the simplest case for arrangements of rational sections over curves . notice that @xmath145 \big),\ ] ] where @xmath146 $ ] are coordinates for @xmath0 . [ e12 ] in the next sections , we will classify all primitive arrangements , and some distinguished subclasses which are defined through intersection properties of their members . we now look at these intersections . in what follows , until the end of this section , we fix the data @xmath129 . let @xmath147 be an arrangement of sections in @xmath80 . let @xmath136 be a point in @xmath80 , and let @xmath148 be local equations defining @xmath149 at @xmath136 . as in ( * ? ? ? * v , section 1 ) , we define the * intersection multiplicity * @xmath150 of @xmath86 and @xmath87 at @xmath136 to be the length of @xmath151 . if @xmath136 is not in @xmath86 or @xmath87 , then @xmath152 . notice that , since @xmath81 , we have @xmath153 . we distinguish two classes of arrangements : * we say that @xmath5 is * transversal * if for any @xmath154 and any point @xmath136 in @xmath155 , there is @xmath156 such that @xmath157 . the set of isomorphism classes of primitive transversal arrangements is denoted by @xmath158 . * we say that @xmath5 is * * simple crossing * * if for any @xmath154 and any point @xmath136 in @xmath159 , we have @xmath160 . this is , the members of the arrangement are pairwise transversal . the set of isomorphism classes of primitive simple crossing arrangements is denoted by @xmath161 . [ d14 ] in @xmath162 above , we have the requirement @xmath163 . this implies @xmath164 , and so the definition is symmetric on @xmath165 . to see this , let @xmath166 be the blow - up at @xmath136 . let @xmath167 be the strict transforms of @xmath168 , so that @xmath169 , for @xmath170 . here @xmath171 is the exceptional curve of @xmath172 . in this way , we have @xmath173 since @xmath168 is nonsingular at @xmath136 . since @xmath172 is an isomorphism outside of @xmath171 , we have that @xmath174 , where @xmath175 . if @xmath176 , then @xmath177 , and so @xmath178 . one proves the general assertion by induction on @xmath150 . [ r11 ] this gives the stratification @xmath179 notice that for line arrangements @xmath180 , but already for @xmath181 we have different sets , as the next example shows . consider collections of curves in @xmath0 given by @xmath182 , as shown in figure [ f1 ] . here , @xmath183 is a conic , and @xmath184 are lines . for distinct @xmath83 s , we have different intersections among @xmath185 s . each @xmath186 has a marked point @xmath136 in @xmath183 . out of these configurations , we produce three arrangements of sections in @xmath187 . we blow up @xmath136 , and then we perform an elementary transformation at @xmath188 , which is the intersection of the strict transform of @xmath183 with the exceptional divisor @xmath171 . then , we have an arrangement of sections @xmath189 in @xmath187 , where @xmath87 corresponds to the strict transform of @xmath185 . which produce arrangements in @xmath187.,width=415 ] any possible morphism of arrangements , from @xmath190 to some @xmath191 , would have @xmath192 as target , and the degree of @xmath38 would have to be @xmath193 . for @xmath194 , we have @xmath195 points in @xmath187 where exactly two sections intersect , and @xmath196 where exactly @xmath22 intersect , so @xmath194 is impossible as pull - back of @xmath197 sections in @xmath198 . similar arguments apply to @xmath199 and @xmath200 , and so one easily checks that all of them a primitive . notice that @xmath200 is simple crossing , @xmath199 is a transversal , and @xmath194 is neither , so @xmath201 . [ e14 ] .,width=377 ] assume @xmath35 has characteristic @xmath202 . let @xmath203 , and consider the @xmath35-linear frobenius morphism @xmath204 @xcite , so @xmath12 and @xmath205 are isomorphic as abstract curves . let @xmath206 be the pull - back arrangement by @xmath38 , as in definition [ d12 ] . then , @xmath207 , and for any two members @xmath149 we have @xmath208 when @xmath209 . the simple crossing arrangement @xmath100 is transformed into an arrangement @xmath5 where any two members are tangent at @xmath210 points , each of order @xmath211 . let @xmath73 be a integer . we denote by @xmath3 the moduli space of @xmath212-pointed stable curves of genus zero @xcite . this is a smooth rational projective variety of dimension @xmath213 . the open set @xmath4 parametrizes configurations of @xmath17 distinct labeled points in @xmath46 . the boundary @xmath214 is formed by the following prime divisors : for each subset @xmath215 with @xmath216 and @xmath217 , we let @xmath218 be the divisor whose generic element is a curve with two components : the points marked by @xmath219 in one , and the points marked by @xmath220 in the other . hence @xmath221 , and usually we will assume @xmath222 to avoid repetitions . these divisors are smooth , and @xmath223 is a simple normal crossing . the variety @xmath3 represents a fine moduli space , proper and smooth over @xmath224 . for @xmath225 , the @xmath83-th forgetful map @xmath226 which forgets the @xmath83-th marked point and stabilizes , gives a universal family . we will mainly consider @xmath227 it has @xmath17 distinguished sections @xmath228 , producing the markings on the parametrized curves . let @xmath6 be a nonsingular projective variety , and let @xmath94 be a nonsingular divisor in @xmath6 . let @xmath229 be a curve in @xmath6 . we say that @xmath229 is * transversal * to @xmath94 if locally at any @xmath230 , the curve @xmath229 can be factored in @xmath231 distinct local irreducible curves ( branches of @xmath229 in @xmath232 ) so that @xmath233 for every @xmath83 . if @xmath94 is a sum of nonsingular divisors @xmath234 , we say that @xmath229 is transversal to @xmath94 if it is to each @xmath234 . [ d21 ] below a well - known property for stable families , coming from the construction of @xmath3 . let @xmath235 be a @xmath35-point in @xmath3 . let @xmath229 be a local curve passing through @xmath235 , i.e. , @xmath229 is a irreducible curve defined by functions in @xmath236 $ ] . assume @xmath237 defines @xmath238 , so @xmath239 , and that @xmath240 is not identically zero for all @xmath241 . consider the commutative diagram @xmath242_{\rho } \ar[r ] & \overline{m}_{0,d+2 } \ar[d]_{\pi_{d+2 } } \\ \overline{b } \ar[r]^{\iota } & \overline{m}_{0,d+1 } } \ ] ] where @xmath243 is the composition of the inclusion of @xmath229 with its normalization , so @xmath244 is the normalization of @xmath229 , and @xmath245 is defined by base change . then , the surface @xmath245 is normal , and can only have singularities of the form @xmath246/(uv - t^m)\ ] ] at the nodes of the singular fiber , for some @xmath247 . moreover , the surface @xmath245 is nonsingular if and only if @xmath229 is transversal to @xmath248 . [ l21 ] a brief outline of the proof . let @xmath249 be the corresponding stable curve over @xmath250 . consider the deformation of @xmath6 as described in @xcite . at a nonsingular point of @xmath6 , we have a nonsingular point for @xmath245 , so we pay only attention to the nodes of @xmath6 . let @xmath251 be a node of @xmath6 , corresponding to the intersection of @xmath229 with @xmath252 for some @xmath253 , i.e. , the node @xmath251 splits @xmath254 in two subsets @xmath219 and @xmath220 , and @xmath255 corresponds to @xmath256 . at the corresponding point @xmath251 in @xmath257 ( over @xmath35 ) , the local rings and the universal map can be written in the projection form @xmath258 \rightarrow \widehat{\o}_{y,\overline{m}_{0,d+2}}= \k[[u_i , v_i , t_1,t_2,\ldots , t_{d-2}]]/(u_iv_i - t_i),\ ] ] for suitable variables @xmath259 . now , the composition of the inclusion of @xmath229 with its normalization @xmath243 has the form @xmath260 for some units @xmath261 s and a local parameter @xmath262 on @xmath244 . this is because @xmath229 is not in @xmath263 for all @xmath264 . hence , @xmath265 for @xmath266 . since @xmath245 is defined through the base change by @xmath243 , we have the isomorphism @xmath267/(u_i v_i - t^{m_i}),\ ] ] and so , @xmath245 is nonsingular iff @xmath268 for all @xmath269 , i.e. , transversal to @xmath248 . the moduli spaces @xmath3 have a beautiful construction , due to kapranov @xcite , as iterated blow - ups of @xmath2 ( see below ) . it follows that curves in @xmath4 are strict transforms of curves in @xmath2 , which are not contained in a certain fixed hyperplane arrangement @xmath270 . the following description of these spaces can be found in @xcite . a * veronese curve * is a rational normal curve of degree @xmath213 in @xmath2 , i.e. , a curve projectively equivalent to @xmath46 in its veronese embedding . [ d22 ] it is a classical fact that any @xmath17 points in @xmath2 in general position lie on a unique veronese curve . the points @xmath271 are said to be in general position if no @xmath272 of them lie in a hyperplane . the main theorem in @xcite says that the set of veronese curves in @xmath2 and its closure are isomorphic to @xmath273 and @xmath274 respectively . ( kapranov @xcite ) take @xmath1 points @xmath275 of projective space @xmath2 which are in general position . let @xmath276 be the space of all veronese curves in @xmath2 through these @xmath1 points @xmath277 . consider it as a subvariety in the hilbert scheme @xmath278 parametrizing all subschemes on @xmath2 . then , * we have @xmath279 . * if @xmath280 is the closure of @xmath276 in @xmath278 , then @xmath281 . the subschemes representing limit positions of curves from @xmath276 are , considered together with @xmath277 , stable @xmath1-pointed curves of genus @xmath282 , which represent the corresponding points of @xmath274 . * the analogs of statements ( a ) and ( b ) hold also for chow variety instead of hilbert scheme . [ t21 ] ( kapranov @xcite ) choose @xmath1 general points @xmath275 in @xmath2 . the variety @xmath3 can be obtained from @xmath2 by a series of blow - ups of all the projective spaces spanned by @xmath277 . the order of these blow ups can be taken as follows : * points @xmath283 and all the projective subspaces spanned by them in order of the increasing dimension ; * the point @xmath284 , all the lines @xmath285 and subspaces spanned by them in order of the increasing dimension ; * the line @xmath286 , the planes @xmath287 , @xmath288 and all subspaces spanned by them in order of the increasing dimension , etc , etc . [ t22 ] let us denote the kapranov s map in theorem [ t22 ] by @xmath289 some conventions and notations for the rest of the paper . let us fix @xmath1 points in general position in @xmath2 . we take @xmath290 $ ] , @xmath291 $ ] , ... , @xmath292 $ ] and @xmath293 $ ] . the symbol @xmath294 denotes the projective linear space spanned by the points @xmath295 . let @xmath296 where @xmath297 and @xmath298 are distinct numbers , and let @xmath299 be the union of all the hyperplanes @xmath300 . hence , @xmath301 \in \p^{d-2 } : x_i = x_j \}$ ] for @xmath302 , @xmath303 \in \p^{d-2 } : x_i=0 \}$ ] and @xmath304 \in \p^{d-2 } : \ x_1 x_2\cdots x_{d-1 } \prod_{i < j } ( x_j - x_i)=0 \}.$ ] for @xmath305 , theorem [ t22 ] says that the map @xmath306 is the blow - up of @xmath0 at the points @xmath307 $ ] , @xmath308 $ ] , @xmath309 $ ] , and @xmath310 $ ] . the hyperplane arrangement @xmath311 is given by the complete quadrilateral @xmath312 the universal map @xmath313 is induced by the pencil of conics ( veronese curves in @xmath0 ) containing @xmath314 , @xmath315 , @xmath316 , and @xmath317 . let @xmath229 be an irreducible projective curve in @xmath3 with @xmath318 . by using kapranov s map @xmath319 , this is the same as giving an irreducible projective curve @xmath10 in @xmath2 not contained in @xmath270 . the proper transform of @xmath10 under @xmath320 is @xmath229 . consider the base change diagram @xmath242_{\rho } \ar[r ] & \overline{m}_{0,d+2 } \ar[d]_{\pi_{d+2 } } \\ \overline{b } \ar[r]^{\iota } & \overline{m}_{0,d+1 } } \ ] ] where @xmath243 is the composition of the inclusion of @xmath229 with its normalization . let us denote @xmath321 . notice that the distinguished sections @xmath322 of @xmath323 induce @xmath17 sections @xmath324 for @xmath325 . also , by lemma [ l21 ] , the surface @xmath245 is a normal projective surface with only canonical singularities of type @xmath326 for various integers @xmath247 , and only at nodes of singular fibers . we now resolve these singularities minimally to obtain a fibration @xmath327 , so that @xmath328 is nonsingular . notice that @xmath329 has only reduced trees of @xmath46 s as fibers , and it has @xmath17 distinguished sections . let @xmath58 be a singular fiber of @xmath329 . consider the curves @xmath171 in @xmath58 with @xmath330 , and which do not intersect the @xmath212-th section ( the proper transform of @xmath331 ) . then , the @xmath171 s are disjoint with self - intersection @xmath332 . we now blow down all of these @xmath171 s to obtain a new fibration over @xmath12 with @xmath17 distinguished sections , and reduced trees of @xmath46 s as fibers . if there is a singular fiber @xmath58 , we repeat the previous procedure . after finitely many steps , this stops in a fibration @xmath333 with nonsingular fibers , and @xmath17 labeled sections @xmath334 , where @xmath335 is the proper transform of @xmath86 . the fibration @xmath333 is isomorphic over @xmath12 to @xmath336 with @xmath337 , and so @xmath338 . the labeled set @xmath339 is a primitive arrangement of @xmath1 sections . [ p31 ] by ( * ? ? ? * v , proposition 2.8 ) , the ruled surface @xmath340 is isomorphic over @xmath12 to @xmath341 , where @xmath342 is a rank two locally free sheaf on @xmath12 with the property that @xmath343 but for all invertible sheaves @xmath344 on @xmath12 with @xmath345 , we have @xmath346 . so , we assume @xmath347 . since , by construction , @xmath348 for all @xmath349 , we have that @xmath342 is decomposable , say @xmath350 , where @xmath13 is unique up to isomorphism . moreover , @xmath351 for all @xmath349 . in particular , @xmath352 for all @xmath353 and @xmath354 . notice that @xmath355 . hence , @xmath356 is an arrangement of @xmath1 sections of @xmath336 . observe that @xmath5 is primitive because it comes from the normalization of a curve in @xmath3 . assume there is a morphism of arrangements from @xmath5 to @xmath100 , with data @xmath130 and map @xmath113 ( see definition [ d12 ] ) . then , our map @xmath357 would factor through @xmath38 , induced by the natural map @xmath358 . this is possible only if @xmath359 , because @xmath12 is the normalization of @xmath229 , and so @xmath5 and @xmath100 are isomorphic arrangements . let @xmath92 be a point in @xmath12 , and consider the fiber @xmath56 of @xmath52 . let @xmath360 be distinct sections which intersect at a point @xmath136 in @xmath56 . then , through the description in lemma [ l21 ] , it is not hard to see that @xmath361 where @xmath362 is the corresponding local branch of @xmath229 at @xmath363 . let @xmath364 . let @xmath365 be the section corresponding to @xmath87 . then , @xmath366 and so , because of the previous formula , @xmath367 . now , by kapranov s description in theorem [ t22 ] , we have @xmath368 and so @xmath337 . this comes from the pull - back of the hyperplane @xmath300 . by the projection formula , we have @xmath369 fix an integer @xmath73 , and an algebraically closed field @xmath35 . we have @xmath370 where the disjoint union is over all nonsingular projective curves @xmath12 and line bundles @xmath13 on @xmath12 ( both up to isomorphism ) . this equality gives a bijection between @xmath131 and the set of irreducible projective curves @xmath229 in @xmath3 with @xmath371 , whose normalization is @xmath12 and @xmath372 , where @xmath373 is the composition of the inclusion of @xmath229 and its normalization . [ t41 ] let @xmath229 be an irreducible curve in @xmath3 with @xmath374 . by proposition [ p31 ] , @xmath229 produces a unique element in @xmath375 , where @xmath243 is the composition of the inclusion of @xmath229 and its normalization . in this way , we only need to prove that given @xmath206 , there is an irreducible curve @xmath229 in @xmath3 intersecting @xmath4 so that @xmath5 is induced by @xmath229 as in proposition [ p31 ] . let @xmath376 be a primitive arrangement of @xmath1 sections of @xmath377 . the section @xmath54 is denoted by @xmath378 . we repeatedly perform blow - ups at the intersections of the sections @xmath86 and their proper transforms , until they are all disjoint . we do this in a minimal way , that is , given a @xmath72-curve in a fiber , its blow - down produces an intersection of the distinguished sections . the corresponding fibration @xmath379 has @xmath212-pointed semi - stable genus zero curves as fibers . the @xmath17 markings are produced by intersecting the proper transforms of the sections @xmath86 s with the fibers . they may fail to be stable exactly because of the presence of fibers with @xmath46 s having no markings , and intersecting the rest of the fiber at two points . these components form chains of @xmath46 s , which we blow down to obtain a @xmath212-pointed stable family of genus zero curves @xmath380 . therefore , we have a commutative diagram @xmath381 \ar[r ] & \overline{m}_{0,d+2 } \ar[d]_{\pi_{d+2 } } \\ c \ar[r ] & \overline{m}_{0,d+1 } } \ ] ] so that @xmath382 . notice that the image of @xmath12 is a curve @xmath229 because @xmath383 ( so not a point ) , and @xmath229 intersects @xmath4 . let @xmath244 be the normalization of @xmath229 , and let @xmath384 be the corresponding map . then , the diagram above factors as @xmath381 \ar[r ] & r \ar[d ] \ar[r ] & \overline{m}_{0,d+2 } \ar[d]_{\pi_{d+2 } } \\ c \ar[r]^{f } & \overline{b } \ar[r]^{\iota } & \overline{m}_{0,d+1 } } \ ] ] where @xmath245 is given by pull - back , and @xmath385 . let @xmath328 be the minimal resolution of the singularities of @xmath245 . let us finally consider the commutative diagram @xmath386 \ar[r ] & \widetilde{r } \ar[d ] \\ c \ar[r]^f & \overline{b } } \ ] ] where @xmath387 ( it may be singular ) . the pull - back of the @xmath17 distinguished sections are the @xmath17 distinguished sections of @xmath388 . we now inductively blow - down all @xmath72-curves on the fibers of @xmath389 in the following way . let @xmath390 be the fibration produced in the @xmath83-th step , where @xmath391 . then , @xmath392 . we obtain the fibration @xmath393 through the commutative diagram below.@xmath394 \ar[dl ] & & & \\ t_{i+1 } \ar@/_/[rrd]_{\simeq } & t_i = c \times_{\overline{b } } r_i \ar[rrr ] \ar[rd ] \ar[ddr ] & & & r_i \ar[dl ] \ar[ddl ] \\ & & c \times_{\overline{b } } r_{i+1 } \ar[r ] \ar[d ] & r_{i+1 } \ar[d ] & \\ & & c \ar[r]^f & \overline{b } & } \ ] ] let @xmath171 be a @xmath72-curve in a fiber of @xmath390 , and let @xmath136 be the point of intersection with the rest of the fiber . notice that at least two distinguished sections @xmath395 and @xmath396 intersect @xmath171 ( not at @xmath136 , of course ) . let @xmath397 be the blow - down of @xmath171 , and @xmath398 be the corresponding fibration . let @xmath399 be the pre - image of @xmath136 and @xmath58 the pre - image of @xmath171 in @xmath400 . notice that @xmath400 may be singular at @xmath399 , say with a singularity of type @xmath401 . if @xmath402 we resolve @xmath399 to get @xmath403 . then we define @xmath404 to be the blow - down of the total transform of @xmath58 in @xmath403 ( this is a chain of @xmath72-curves ) . let us consider @xmath136 in @xmath397 , and its pre - image in @xmath405 , say @xmath406 . now , @xmath407 is nonsingular at @xmath406 , and there is a morphism @xmath408 , which is clearly an isomorphism . for @xmath409 , these procedure is what we have in section [ s3 ] . when it stops , say at the @xmath247-th step , we have @xmath410 with @xmath17 distinguished sections , and @xmath411 nonsingular @xmath46-bundle over @xmath12 . moreover , because of the construction of @xmath400 s , we have @xmath412 , and the arrangement @xmath5 is the pull - back of the one in @xmath413 . so this is a morphism of arrangements as in definition [ d12 ] . but @xmath5 primitive implies @xmath414 , and so @xmath415 . for example , one has @xmath416 which corresponds to the unique curve in @xmath417,[1:1],[1:0]\}$ ] . an immediate corollary is the following . given a nonsingular projective curve @xmath12 and a line bundle @xmath13 on @xmath12 , the kapranov s map @xmath418 gives a one - to - one correspondence between elements of @xmath131 and irreducible projective curves @xmath10 in @xmath2 not contained in @xmath270 such that @xmath419 and @xmath420 . in particular , @xmath421 . [ c41 ] @xmath422 . [ c42 ] a curve of degree one in @xmath2 is a line . @xmath423 . [ c43 ] an irreducible curve of degree two is a conic . the next two corollaries identify precisely the two distinguished classes of arrangements in definition [ d14 ] . @xmath424 where the disjoint union is over all nonsingular projective curves @xmath12 and line bundles @xmath13 on @xmath12 , both up to isomorphism . [ c44 ] let @xmath425 . let @xmath136 be a singular point of the reducible curve defined by @xmath5 in @xmath80 . let @xmath56 be the fiber containing @xmath136 . hence , since @xmath5 satisfies ( t ) in definition [ d14 ] , there are two transversal sections @xmath149 containing @xmath136 , i.e. , @xmath426 . consider the blow - up at @xmath136 , @xmath427 , and let @xmath171 be the exceptional curve . then , @xmath171 has at least three special distinct points : the intersections with @xmath428 , @xmath429 , and @xmath430 ( corresponding proper transforms ) . now , it is clear that the final * stable * fibration @xmath431 produced from @xmath5 has the proper transform of @xmath171 as a component of the fiber over @xmath92 . let @xmath432 be the collection of proper transforms of @xmath86 s in @xmath433 . then , @xmath434 satisfies property ( t ) in definition [ d14 ] ( extending naturally this definition ) . so , we repeat the blow - ups until all sections are disjoint ( in a minimal way ) to obtain the stable fibration @xmath431 , where no blow - downs are needed . since @xmath245 is a nonsingular surface , the curve @xmath435 is transversal to @xmath248 by lemma [ l21 ] . now assume @xmath5 is not in @xmath158 . then , there are indices @xmath165 and a point @xmath209 such that @xmath436 . let @xmath437 , so @xmath438 . we blow up @xmath439 times the corresponding point in @xmath440 for the successive proper transforms of @xmath86 and @xmath87 . let @xmath6 be the resulting surface , and @xmath441 . let @xmath171 be the exceptional curve of the blow - up at @xmath442 . then , @xmath171 has only two special points : the intersection with the rest of the fiber and with the section @xmath429 . notice that @xmath443 at @xmath171 . so , in the process to obtain the corresponding stable fibration @xmath431 , we need to blow up again at @xmath440 , and so at the end the proper transform of @xmath171 will have to be blown down ( in order to have a stable fibration ) . therefore , @xmath245 is singular , and by lemma [ l21 ] , @xmath435 is not transversal to @xmath248 . by theorem [ t41 ] , we have checked all irreducible curves in @xmath4 . @xmath444 where the disjoint union is over all nonsingular projective curves @xmath12 and line bundles @xmath13 on @xmath12 , both up to isomorphism . [ c45 ] let @xmath445 , and consider its stable fibration @xmath446 . we know that the image of @xmath12 in @xmath3 is transversal to @xmath248 by the previous corollary . let @xmath57 be a point whose fiber is singular . then there exists an element in @xmath143 that produces the same fiber . by corollary [ c42 ] , the set @xmath143 is in one - to - one correspondence with lines in @xmath2 not in @xmath270 . therefore , there is a line in @xmath2 not in @xmath270 passing through @xmath447 . this implies that the image of @xmath435 under @xmath448 is transversal to @xmath270 . the converse is clear using the same correspondence with lines . in the previous section we classified all arrangements of @xmath1 sections ( and two distinguished subclasses ) . they are in one - to - one correspondence with curves in @xmath2 outside of a certain fixed hyperplane arrangement @xmath270 ( theorem [ t41 ] ) . in @xcite , we used this correspondence to explicitly find new special line arrangements in @xmath0 . for this , we computed the corresponding line as in corollary [ c42 ] . in general , it is hard to present a curve in @xmath2 in the form we need to construct its corresponding arrangement . in this brief section , we show a simple way to produce arrangements via irreducible curves in @xmath0 . this is based on ( * ? ? ? * section 7 ) , where castravet and tevelev describe how to cover @xmath4 with blow - ups of @xmath0 at @xmath17 points . * proposition 7.3 ) suppose @xmath449 are distinct points in @xmath0 , and let @xmath450 be the complement to the union of lines containing at least two of them . the morphism @xmath451 obtained by projecting @xmath449 from points of @xmath395 extends to the morphism @xmath452 if there is no ( probably reducible ) conic through @xmath449 then @xmath453 is a closed embedding . in this case the boundary divisors @xmath256 of @xmath3 pull - back as follows : for each line @xmath454 , we have @xmath455 ( the proper transform of @xmath456 ) and ( assuming @xmath457 ) , @xmath458 , where @xmath459 and @xmath460 is the exceptional divisor over @xmath461 . other boundary divisors pull - back trivially . [ p51 ] in this way , we have @xmath462 where @xmath463 is the class of a general line in @xmath0 , @xmath464 is the number of lines in @xmath0 passing through @xmath465 and some other @xmath466 , and @xmath467 if there is a @xmath461 in @xmath468 @xmath469 or @xmath470 otherwise . hence , the image of @xmath471 under @xmath472 is a surface @xmath473 in @xmath2 of degree @xmath474 , and so @xmath475 therefore , @xmath473 is a surface of minimal degree in some @xmath476 . thus @xmath473 is either a rational normal scroll in @xmath477 or the veronese of @xmath0 in @xmath478 . moreover , @xmath473 is smooth . one can check that @xmath320 blows down certain @xmath1 @xmath72-curves in @xmath471 ( proper transforms of lines @xmath479 with @xmath470 , and @xmath480 with @xmath467 ) having as result a hirzebruch surface @xmath481 , where @xmath247 depends on the configuration of points @xmath482 such that @xmath470 . given @xmath449 points in @xmath0 , with no ( probably reducible ) conic through them , we consider an irreducible plane curve @xmath483 not included in the union of lines containing @xmath449 . then , by proposition [ p51 ] , we have the inclusion @xmath484 and so a primitive arrangement @xmath5 in @xmath485 for some line bundle @xmath13 , by theorem [ t41 ] . the line bundle @xmath13 depends on the specific configuration @xmath449 and the position of the curve @xmath483 with respect to these points . proposition [ p51 ] gives a way to precisely see all possible intersections of @xmath483 with @xmath248 , and so this procedure indeed gives an explicit description of @xmath5 . .,width=453 ] let @xmath486 . consider the dual hesse arrangement of @xmath487 lines in @xmath488 : @xmath489 it has @xmath490 triple points and no other singularities . we label these points as @xmath491 $ ] , @xmath492 $ ] , @xmath493 $ ] , @xmath494 $ ] , @xmath495 $ ] , @xmath496 $ ] , @xmath497 $ ] , @xmath498 $ ] , @xmath499 $ ] , @xmath500 $ ] , @xmath501 $ ] , and @xmath502 $ ] . consider the unique conic @xmath483 passing through @xmath503 . it is given by the equation @xmath504 . by proposition [ p51 ] , we have @xmath505 , embedding the proper transform @xmath506 of @xmath483 . in @xcite , it is proved that @xmath506 is a rigid curve in @xmath507 . by theorem [ t41 ] , this curve defines a primitive arrangement of @xmath508 curves @xmath5 . to actually exhibit @xmath5 , we need to check all intersections between @xmath483 and all the lines passing through pairs of points @xmath482 ( so , more than the ones in the dual hesse arrangement ) . after that , it is easy to draw a picture of the arrangement . in figure [ f3 ] , we show all the singular fibers of the corresponding stable fibration . notice that the arrangement belongs to @xmath509 . for another model , we perform in @xmath510 two elementary transformations on the fibers @xmath511 and @xmath512 by blowing up the corresponding singular points in @xmath5 . then , we end up in @xmath71 where the @xmath72-curve is the proper transform of @xmath513 . after blowing it down , we obtain a very special arrangement of @xmath514 lines and @xmath197 conics in @xmath0 . fix the data @xmath129 over @xmath515 as in section [ s1 ] . we now study properties of certain log surfaces associated to arrangements of sections @xmath5 . first , we associate to @xmath5 an extended arrangement @xmath20 , and partially extended arrangements @xmath516 . consider the arrangement of sections @xmath5 as a ( reducible singular ) curve , we denote its set of singular points by @xmath517 . let @xmath518 be the fibers of @xmath377 which contain points in @xmath517 . then , the * extended * arrangement @xmath20 associated to @xmath5 is @xmath519 let @xmath520 be an integer . let @xmath521 be a subset of @xmath522 such that for any @xmath523 and any point @xmath136 in @xmath524 , there are two sections in @xmath5 intersecting at @xmath136 with distinct tangent directions . then , a * partially extended * arrangement @xmath525 associated to @xmath5 is @xmath526 the numeration of the fibers will be irrelevant . [ d61 ] as before , we perform blow - ups at the points in @xmath517 ( and infinitely near points above them ) to separate all sections @xmath86 s in a minimal way ( as in the proof of theorem [ t41 ] ) . this is described by a chain of blow - ups @xmath527 whose composition is denoted by @xmath528 . the map @xmath172 gives the * minimal log resolution * of @xmath5 , and produces the semi - stable fibration of @xmath212-pointed genus zero curves @xmath529 . the @xmath530 is a simple normal crossing divisor . let @xmath531 be the number of sections in the proper transform of @xmath5 passing through @xmath277 right before blowing up @xmath277 ( the center of the blowing up @xmath532 ) . we define @xmath533 the divisor @xmath534 is the minimal log resolution of @xmath20 , but @xmath535 may not be minimal , since we may need to blow down @xmath72-curves coming from some nodes in @xmath5 . the divisors @xmath534 and @xmath535 are denoted by @xmath536 and @xmath537 respectively . the arrangement @xmath536 may be seen as defined by the intersection of the boundary @xmath248 in @xmath257 with the surface @xmath245 , where @xmath446 is the stable fibration of @xmath212-pointed curves of genus zero induced by @xmath5 . we now follow the exposition of log surfaces as in ( * ? ? ? * section 2 ) , which is due to iitaka , and the references given there . we are interested in the log surfaces @xmath538 and @xmath539 , and their * * log chern classes * * and the minimal log resolution of @xmath21 are the same ( see for example ( * ? ? ? * proposition 2.4 ) ) . ] @xmath540 where @xmath541 , @xmath542 are the dual of the corresponding sheaves of log differentials ( see ( * ? ? ? * def.2.2 ) ) , and @xmath543 . let @xmath5 be an arrangement of sections with data @xmath129 , @xmath544 and @xmath545 . then , @xmath546 with @xmath547 as in definition [ d61 ] . [ p61 ] in general , if @xmath548 is a log surface ( as in ( * ? ? * section 2 ) ) , the log chern numbers are ( see ( * ? ? ? * proposition 2.4 ) ) @xmath549 and @xmath550 , where @xmath551 is the number of nodes of the curve @xmath94 . in our case , @xmath552 and @xmath553 . we will compute these numbers recursively . let @xmath554 be the composition of the blow - ups @xmath555 . define @xmath556 for @xmath557 , where @xmath558 is the prime decomposition of @xmath559 . then , one can check that @xmath560 and @xmath561 . therefore , @xmath562 . on the other hand , by the formula for @xmath563 above , we have @xmath564 let @xmath356 as above . then , @xmath565 , @xmath566 , and @xmath567 . the log canonical divisor @xmath568 is big and nef , and so the surface @xmath569 is of log general type . when @xmath37 , we have the ( strict ) log miyaoka - yau inequality @xmath570 and so @xmath571 . [ c61 ] the second log chern number @xmath572 is positive because @xmath573 . for the other one , take @xmath574 . then , by looking at the intersections of @xmath575 with @xmath576 in @xmath80 , one sees that @xmath577 . the inequality is strict because @xmath578 . therefore , by proposition [ p61 ] , one has @xmath579 . by the same reason , @xmath580 let @xmath581 and write its prime decomposition @xmath582 where @xmath429 is the proper transform of @xmath86 under @xmath583 , and @xmath584 s are the rest . since @xmath585 , we denote @xmath586 and @xmath587 where @xmath58 is a general fiber of @xmath588 . then , @xmath589 for some @xmath590 . but , for any fixed @xmath591 , @xmath592 where @xmath593 , and @xmath594 . thus , say for @xmath595 , we have @xmath596 which means that the log canonical class is numerically equivalent to an effective divisor . moreover , @xmath597 for all @xmath83 , and @xmath598 , and so @xmath599 is nef . this plus the fact @xmath600 implies that @xmath601 is big , and so @xmath602 is of log general type . when @xmath37 , we use sakai s theorem ( * ? ? ? * theorem 7.6 ) to conclude @xmath603 . notice that the curve @xmath94 is semi - stable and has no exceptional curves with respect to @xmath94 , as defined in @xcite . for strictness , we apply lemma [ apl ] . from the previous proof , it is easy to see that the log canonical class is ample if and only if @xmath5 is transversal ( definition [ d14 ] ) . one just applies the nakai - moishezon criterion @xcite . [ o61 ] let @xmath5 be an arrangement of sections with data @xmath129 , and let @xmath21 be a partially extended arrangement with @xmath604 for some @xmath605 ( see definition [ d61 ] ) . let @xmath606 and @xmath607 . then , @xmath608 [ p62 ] the result follows directly from the formulas in ( * ? ? ? * proposition 2.4 ) . let @xmath5 be as above . then , @xmath609 , @xmath610 . the corresponding log canonical divisor is big and nef , and so the log surface defined by @xmath537 is of log general type . when @xmath37 , we again have ( strict ) log miyaoka - yau inequality @xmath611 . [ c62 ] notice that @xmath612 , and so @xmath613 clearly , we have @xmath614 , and so @xmath615 . then , since @xmath616 , we have @xmath617 @xmath618 @xmath619 we prove nefness and bigness as we did in corollary [ c61 ] . it is enough to do it in @xmath620 . let @xmath621 and write its prime decomposition @xmath622 where @xmath429 is the proper transform of @xmath86 under @xmath623 , and @xmath584 s are the rest . since @xmath585 , we denote @xmath624 and @xmath587 where @xmath58 is a general fiber of @xmath377 . then , @xmath625 for some @xmath590 . but , for any fixed @xmath591 , @xmath626 where @xmath627 , and @xmath628 ( see definition [ d61 ] ) . thus , say for @xmath595 , we have @xmath629 which means that the log canonical class is numerically equivalent to an effective divisor . moreover , @xmath630 for all @xmath83 , and @xmath631 ( this is by definition of partially extended arrangement ) , and so @xmath599 is nef . this plus the fact @xmath632 implies that @xmath601 is big , and so @xmath633 is of log general type . when @xmath37 and by sakai s theorem ( * ? ? ? * theorem 7.6 ) , we have @xmath634 . for strictness , we apply again lemma [ apl ] . we use example [ e51 ] to show differences in @xmath635 between extended and partially extended arrangements . let @xmath5 be the arrangement in example [ e51 ] . for the partially extended arrangements , let @xmath636 be the set of fibers we take out from @xmath20 ( see definition [ d61 ] ) . we label fibers according to figure [ f3 ] . then we have the following table . [ cols="^,^,^,^,^,^,^",options="header " , ] notice that @xmath637 has no @xmath638 " in case v because of our restriction on tangent directions at the singular points of @xmath21 ( definition [ d61 ] ) . below we estimate for each type the number of solutions @xmath639 of @xmath640 producing a bad multiplicity @xmath641 . we do it case by case . ( * type i ) * this is a node in @xmath155 ( possible only when @xmath642 ) . since @xmath640 is a weighted partition of @xmath211 , we can use the estimate in ( * ? ? ? * proof of theorem 6.1 ( 1 ) ) , and so there exists a positive number @xmath643 ( independent of @xmath211 ) such that @xmath644 * ( type ii ) * this is a node in @xmath645 with @xmath646 . then again , we apply what we did in ( * ? ? ? * proof of theorem 6.1 ) , to obtain the same estimate as above . * ( type iii ) * this is a node in @xmath647 . since @xmath648 , then we want @xmath649 mod @xmath211 , so @xmath650 . but this is again as in ( * ? ? ? * theorem 6.1 ) , and we have the same previous estimate . notice that it works because @xmath651 . * ( type iv ) * this is a node between @xmath652 , @xmath653 , and a exceptional divisor over the fiber @xmath654 . notice that @xmath525 contains at least two fibers , so @xmath655 . hence we are as in case ( 2 ) in the proof of ( * ? ? ? * theorem 6.1 ) . * ( type v ) * this is the new case , coming from nodes in the resolution of singularities of @xmath5 . it does not involve @xmath656 . the idea is to analyze three equations @xmath657 , @xmath658 , @xmath659 from the equation @xmath640 , and estimate solutions for each . without loss of generality , we rearrange indices so that @xmath660 , for some @xmath661 , and @xmath662 , where @xmath663 , for some @xmath664 , with @xmath665 or @xmath196 , and @xmath666 . notice that @xmath667 for any solution of @xmath640 . we define @xmath668 equation with @xmath669 variables , @xmath670 with @xmath671 variables , and @xmath672 with @xmath673 variables . so , @xmath674 . notice that @xmath482 are numbers varying in the region @xmath675 , since we will look at solutions of @xmath676 from solutions of @xmath640 . say that @xmath641 , which means mod @xmath211 , @xmath677 . of course the set @xmath678 has same size as @xmath679 . we now use repeatedly the fact that the number of nonnegative integer solutions of @xmath680 for coprime @xmath681 s is @xmath682 ( see @xcite ) . let @xmath211 be large enough . given @xmath683 , the number of solutions of @xmath659 is @xmath684 . now , the key observation is that mod @xmath211 we have @xmath685 because @xmath686 is not zero . in this way , we have to choose @xmath687 in a set of size @xmath688 . now we fix @xmath687 and have at most @xmath689 solutions for @xmath658 . after we have solutions for @xmath659 and @xmath658 , we have at most @xmath690 solutions for @xmath657 . putting it all together , @xmath691 but we know that @xmath692 ( * ? ? ? * apendix ) , and that the total number of solutions of @xmath640 is @xmath693 . then , since the number of nodes of @xmath694 is of course independent of @xmath211 , we have proved the existence of good solutions , and that a random one is good with probability tending to @xmath196 as @xmath211 becomes arbitrarily large . now , given good solutions with @xmath211 large , we proceed as at the end of the proof of ( * ? ? ? * theorem 6.1 ) , showing that . this is due to the usual normalization by @xmath490 of a dedekind sum . the correct estimate is @xmath695 , which of course does not affect any asymptotic result . ] @xmath696 this proves the asymptotic result . finally , these surfaces are of general type because of the classification of algebraic surfaces ( see @xcite for any characteristic ) , since we know that @xmath697 and @xmath698 ( corollaries [ c61 ] and [ c62 ] ) . with this theorem , one recovers the log miyaoka - yau inequalities in corollaries [ c61 ] and [ c62 ] , when @xmath37 . we just apply the ( projective ) miyaoka - yau inequality to the surfaces @xmath6 for large primes @xmath211 . [ o71 ] a good looking corollary , consequence of theorem [ t41 ] , section [ s6 ] , and theorem [ t71 ] . let @xmath73 be an integer , and let @xmath10 be an irreducible projective curve in @xmath699 not contained in the hyperplane arrangement @xmath270 ( see above example [ e21 ] ) . then , there exist nonsingular projective surfaces @xmath6 associated to @xmath10 such that @xmath6 is of general type and @xmath700 . [ c71 ] consider the arrangement @xmath5 defined by @xmath10 as in corollary [ c41 ] . then use theorem [ t71 ] for @xmath20 , and corollary [ c61 ] . in addition , one can prove that @xmath701 , where @xmath9 is the normalization of the curve @xmath10 , and @xmath702 denotes the tale fundamental group @xcite . [ o72 ] assume @xmath37 . let @xmath5 be an arrangement of sections of @xmath588 . then , there exist nonsingular projective surfaces @xmath6 of general type such that @xmath703 ( and so of positive index ) having @xmath27 arbitrarily close to @xmath704 . in addition , there is an induced connected fibration @xmath705 which gives an isomorphism : @xmath706 . in this way , alb@xmath31 jac@xmath707 and @xmath33 is the albanese fibration of @xmath6 . [ c72 ] the first part is implied by theorem [ t71 ] for @xmath20 and corollary [ c61 ] . the map @xmath708 in the proof of theorem [ t71 ] is totally ramified along @xmath536 , and @xmath709 is a connected fibration with at least one simply connected fiber and one section in @xmath536 . therefore , the construction induces a connected fibration @xmath710 , and by ( * ? ? ? * proposition 8.3 ) we have @xmath711 . the last part is a simple consequence of albanese maps which applies to any such fibration ( see @xcite for example ) . corollary [ c72 ] is also valid for any @xmath21 except for @xmath712 . if one thinks that the closest @xmath27 is to @xmath22 , the more interesting are the surfaces @xmath6 , then one may consider the construction starting with some @xmath21 ( this is with @xmath642 ) . for line arrangements , this is indeed the case ( see remark [ o64 ] ) . by proposition [ p63 ] , we only need to consider primitive arrangements in order to find upper bounds for chern ratios . [ o73 ] the conic in example [ e51 ] produces an arrangement @xmath713 . in the table of example [ e61 ] , we computed log chern ratios for the extended and some partially extend arrangements induced by @xmath5 . then , by corollary [ c72 ] , there are simply connected nonsingular projective surfaces of general type with chern ratios arbitrarily close to the ones in that table . notice that the highest is attained by a partially extended arrangement , which avoids too many " double points . [ e71 ] assume @xmath35 has positive characteristic @xmath211 . take any @xmath714 for some curve @xmath12 and line bundle @xmath13 . consider the @xmath35-linear frobenius pull - back of @xmath100 composed @xmath715 times , as in remark [ o65 ] . denote the resulting arrangement by @xmath5 . then , by remark [ o65 ] and theorem [ t71 ] , there are nonsingular projective surfaces of general type @xmath6 with @xmath27 arbitrarily close to @xmath716 , and so arbitrarily large . we can prove that @xmath717 ( see @xcite ) . therefore , for any given positive characteristic and nonsingular projective curve @xmath12 , there are nonsingular projective surfaces of general type @xmath6 with @xmath718 and violating any sort of miyaoka - yau inequality . in this section , the ground field is @xmath7 . after fixing @xmath131 , it is clearly of our interest to find optimal upper bounds for @xmath635 for extended and partially extended arrangements ( see remark [ o64 ] ) . arrangements attaining upper bounds should be very special , and they would produce interesting surfaces via theorem [ t71 ] . in this appendix , we show through theorems [ apt ] and [ aptt ] how this question about sharp upper bounds is connected to old questions by lang and others on effective height inequalities @xcite , via an inequality of liu ( * ? ? ? * theorem 0.1 ) . also , in a more general setting , we show a way to obtain strictness for the log inequalities in corollaries [ c61 ] and [ c62 ] , and theorems [ apt ] and [ aptt ] . the next lemma follows from kobayashi @xcite and mok @xcite . by @xcite , @xmath722 big and nef and @xmath721 imply that the universal covering of @xmath723 is the complex two dimensional ball @xmath724 . hence , there exist a discrete group @xmath483 in @xmath725 such that @xmath726 . in particular , @xmath727 has finite volume . notice that @xmath483 is torsion - free since it acts freely on @xmath728 . therefore , by ( * ? ? ? * main theorem ) , there exists a smooth projective mumford compactification @xmath729 of @xmath727 such that @xmath730 is a disjoint union of smooth elliptic curves @xmath480 . in this way , we obtain a birational map @xmath731 . we now resolve this map and get a birational morphism @xmath732 . then , the inverse image @xmath733 of each @xmath480 under @xmath172 dominates @xmath734 , after reordering indices . it is easy to see that @xmath735 . but @xmath733 is a smooth elliptic curve with some finite trees of @xmath46 s attached . given one of these trees , one has a smooth rational curve @xmath58 intersecting @xmath736 at one point . but then @xmath737 . so , there are no trees , and @xmath738 for all @xmath83 . let @xmath739 be a fibration of a smooth projective surface over a smooth projective curve @xmath12 , denote by @xmath38 the genus of the generic ( connected ) fiber of @xmath740 and by @xmath741 the genus of @xmath12 . let @xmath742 be the relative dualizing sheaf . let @xmath743 be @xmath439 mutually disjoint sections of @xmath740 . assume @xmath740 is a semi - stable fibration of @xmath439-pointed curves of genus @xmath38 , marked by these sections . let @xmath744 where @xmath745 are the images of the singular fibers of @xmath740 . the generic fiber has @xmath750 ( @xmath751 denotes the log kodaira dimension ) : @xmath46 minus at least four points or elliptic curve minus at least one point or the rest . now , since @xmath740 is not isotrivial , it has at least @xmath22 singular fibers when @xmath15 ( see @xcite ) , or at least one when @xmath12 is an elliptic curve ( see @xcite ) . so , in any case , the base is of log general type ( on the base we take the log curve @xmath752 ) . by a theorem of kawamata ( * ? ? ? * theorem 11.15 ) , we have for a general @xmath57 @xmath753 and so @xmath754 is of log general type . notice that @xmath94 is a semi - stable curve , just because the fibration is semi - stable , and when @xmath15 we have at least @xmath22 singular fibres . therefore , sakai s theorem @xcite applies , and so @xmath755 . below we show that @xmath722 is nef to obtain the strict inequality . ( case @xmath757 ) : let @xmath758 be the relative minimal model of @xmath740 . let @xmath480 be the exceptional divisors . we can write @xmath759 where for some positive integers @xmath760 s . since @xmath757 , the dualizing sheaf @xmath761 is nef ( due to arakelov ) . if @xmath762 , then @xmath763 is nef as well , and so we check that @xmath764 is nef by intersecting it with components of @xmath94 ( notice that @xmath94 includes @xmath460 s ) . if @xmath765 , we have at least @xmath22 singular fibers , and so we delete @xmath763 using @xmath94 , and again to check nef we intersect @xmath764 with the components of @xmath94 . ( case @xmath748 ) : as in the previous case , we go to @xmath766 . notice that @xmath766 has no multiple fiber . by the canonical bundle formula we have @xcite @xmath767 where @xmath58 is a general fiber of @xmath740 . so , if @xmath762 , we are done by the previous argument . if @xmath765 , we are done by the same argument , since there are at least @xmath22 singular fibers by @xcite . theorem [ apt ] is also valid when @xmath768 ( i.e. , a kodaira fibration ) . it follows from ( * ? ? ? * theorem 0.1 ) . our argument does not prove it . actually , up to the case @xmath769 , theorem [ apt ] is just a small extension of ( * ? ? ? * theorem 0.1 ) ( since we also consider the cases @xmath770 ) , as we now see . we have @xmath771 ( @xmath772 is the euler topological characteristic of @xmath10 ) and the usual formula ( * ? ? ? * lemma vi.4 ) @xmath773 where @xmath58 is a generic fiber of @xmath740 . one sees that @xmath774 . so , @xmath775 obviously @xmath776 . we then square it and see the inequality in ( * ? ? ? * theorem 0.1 ) . in ( * ? ? ? * theorem 0.1 ) , where @xmath777 is assumed , we have that @xmath740 is isotrivial if and only if @xmath778 . but we now show that this corresponds to uninteresting situations . first notice that @xmath764 is nef by the same argument used in ( case @xmath757 ) of theorem [ apt ] . if @xmath779 , then @xmath764 becomes big and nef , and we apply lemma [ apl ] to obtain a contradiction , unless @xmath780 and @xmath781 . but then @xmath782 , which is a contradiction to our assumption @xmath779 . therefore , we are in the trivial case @xmath783 . [ apo1 ] for completeness sake , we explicitly show the connection with height inequalities of algebraic points on curves over function fields . this is another proof of tan s height inequality ( * ? ? ? * theorem a ) . let @xmath739 be a connected fibration as before , denoting by @xmath38 the genus of the generic fiber of @xmath740 and by @xmath741 the genus of @xmath12 . assume that @xmath740 is semi - stable . let @xmath784 be the function field of @xmath12 . for an algebraic point @xmath785 , let @xmath786 be the corresponding horizontal curve ( i.e. multisection ) in @xmath719 . as usual , let @xmath787 be the geometric height and the geometric logarithmic discriminant respectively . the curve @xmath788 is the normalization of @xmath786 , and @xmath58 is a general fiber of @xmath740 . where @xmath793 is the unique semi - stable fibration induced by @xmath38 . notice that @xmath794 contains a section @xmath473 of @xmath793 , by construction . the map @xmath795 is the induced semi - stable fibration with a marked point ( marked by @xmath473 ) . let @xmath796 be the number of singular fibers of @xmath795 . notice that @xmath796 is at most @xmath797 . consider @xmath798 where @xmath799 are the images of the singular fibers of @xmath795 in @xmath788 , and @xmath800 is the strict transform of @xmath473 . we now apply theorem [ apt ] to have @xmath801 . but , one checks that @xmath802 . also , since @xmath793 is semi - stable , we know that @xmath803 , and by the projection formula @xmath804 so , the log inequality above becomes w. p. barth , k. hulek , c. a. m. peters , and a. van de ven . _ compact complex surfaces _ , ergebnisse der mathematik und ihrer grenzgebiete . 3 . folge . , second edition , vol . 4 , springer - verlag , berlin , 2004 . n. mok . _ projective - algebraicity of minimal compactifications of complex - hyperbolic space forms of finite volume _ , pre - print in http://hkumath.hku.hk/@xmath812imr/imrpreprintseries/2009/imr2099-4.pdf ( 2009 ) .	we introduce arrangements of rational sections over curves . they generalize line arrangements on @xmath0 . each arrangement of @xmath1 sections defines a single curve in @xmath2 through the kapranov s construction of @xmath3 . we show a one - to - one correspondence between arrangements of @xmath1 sections and irreducible curves in @xmath4 , giving also correspondences for two distinguished subclasses : transversal and simple crossing . then , we associate to each arrangement @xmath5 ( and so to each irreducible curve in @xmath4 ) several families of nonsingular projective surfaces @xmath6 of general type with chern numbers asymptotically proportional to various log chern numbers defined by @xmath5 . for example , for the main families and over @xmath7 , any such @xmath6 is of positive index and @xmath8 , where @xmath9 is the normalization of @xmath10 . in this way , any rational curve in @xmath4 produces simply connected surfaces with @xmath11 . inequalities like these come from log chern inequalities , which are in general connected to geometric height inequalities ( see appendix ) . along the way , we show examples of tale simply connected surfaces of general type in any characteristic violating any sort of miyaoka - yau inequality . arrangements of rational sections over curves set up a new class of arrangements of curves on algebraic surfaces . given a nonsingular projective curve @xmath12 and an invertible sheaf @xmath13 on @xmath12 , they are defined as finite collections of sections of the corresponding @xmath14-bundle . the simplest example is line arrangements on @xmath0 , where @xmath15 and @xmath16 . in section [ s1 ] , we systematically study arrangements of rational sections over curves . although in somehow they can be managed as line arrangements , the big difference relies on possible tangencies among their curves , introducing more geometric liberties . we partially organize this via transversal and simple crossing arrangements ( definition [ d14 ] ) . throughout sections [ s2 ] , [ s3 ] and [ s4 ] , we show one - to - one correspondences between arrangements of @xmath1 sections and irreducible curves in @xmath4 , the moduli space of curves of genus zero with @xmath17 marked ordered points . this is done for each fixed pair @xmath18 in the general ( theorem [ t41 ] ) , transversal ( corollary [ c44 ] ) , and simple crossing ( corollary [ c45 ] ) cases . because of kapranov s description of @xmath3 @xcite , this produces bijections between arrangements and curves in @xmath2 ( corollary [ c41 ] ) . for instance , arrangements of @xmath1 lines in @xmath0 correspond to lines in @xmath2 ( as in @xcite ) , arrangements of @xmath1 conics in @xmath19 correspond to conics in @xmath2 , etc . to exemplify , we show in section [ s5 ] a way to produce explicit arrangements of sections from irreducible curves in @xmath0 . this is based on ( * ? ? ? * section 7 ) , where the authors show how to cover @xmath4 with blow - ups of @xmath0 at @xmath17 points . we use their rigid conic as concrete example ( see examples [ e51 ] and [ e61 ] ) . given an arrangement of sections @xmath5 , we define two types of arrangements : the extended @xmath20 and some partially extended @xmath21 . their definitions and log properties are exposed in section [ s6 ] . over @xmath7 , they satisfy certain log miyaoka - yau inequalities which are no longer combinatorial as in the case of line arrangements ( remarks [ o64 ] and [ o62 ] ) . for line arrangements we know an optimal log inequality ( hirzebruch - sakai s in remark [ o64 ] ) , but for any other class we only have the coarse bound @xmath22 . arrangements attaining upper bounds should be special , and would produce interesting algebraic surfaces by means of theorem [ t71 ] . we remark that questions about sharp upper bounds of log chern ratios are related to questions on effective height inequalities @xcite ( see appendix , where we slightly extend and give another proof of liu s inequality ( * ? ? ? * theorem 0.1 ) , which naturally implies strict tan s height inequality ( * ? ? * theorem a ) ) . in section [ s7 ] , we associate various families of nonsingular projective surfaces to any given arrangement of sections @xmath5 . these surfaces share the random nature of the surfaces @xmath6 constructed in @xcite , having chern invariants asymptotically proportional to the log chern invariants of @xmath20 and @xmath21 s . in this way , we are able to show for a more general class of arrangements ( and so singularities ) that the behavior of dedekind sums and continued fractions used in @xcite can also be applied . in this paper , any such @xmath6 is of general type and satisfies @xmath23 . putting it all together , and over @xmath7 , we have the following . * theorem . * let @xmath10 be an irreducible curve in @xmath24 not contained in @xmath25 . let @xmath9 be the normalization of @xmath10 . then , there exist nonsingular projective surfaces @xmath6 of general type such that @xmath26 having @xmath27 arbitrarily close to @xmath28 , a well - defined positive rational number depending on @xmath10 and its position in @xmath24 . in addition , there is an induced connected fibration @xmath29 which gives an isomorphism @xmath30 . in this way , alb@xmath31 jac@xmath32 and @xmath33 is the albanese fibration of @xmath6 . with this in hand , we aim to answer the still open question : are there simply connected nonsingular projective surfaces of general type with @xmath34 arbitrarily close to the miyaoka - yau bound @xmath22 ? hence , at least when @xmath10 is a rational curve , it is important for us to know about sharp upper bounds for @xmath28 ( also for @xmath21 , see remark [ o73 ] ) . so far , we only know that this bound is strictly smaller than @xmath22 ( corollaries [ c61 ] and [ c62 ] ) . on the other hand , in positive characteristic , we use our method to produce tale simply connected nonsingular projective surfaces of general type which violate any sort of miyaoka - yau inequality for any given characteristic ( example [ e15 ] , remark [ o65 ] , example [ e72 ] ) . * acknowledgments * : i am grateful to i. dolgachev , s. keel , f. knudsen , k. liu , n. mok , a. momot , j. tevelev , t. weston and s. wong for valuable conversations and e - mail correspondence . i would like to thank the referee for such a complete and helpful review which improved much many parts of this article .
is a transient x - ray binary system located near the galactic center , discovered in august 1989 using the imaging spectrometer aboard the _ mir - kvant _ observatory @xcite . three type i x - ray bursts were observed from the source during 24 1000 s observations spanning 15 days @xcite , which established that the source was a neutron star low mass x - ray binary ( lmxb ) . subsequent rosat observations confirmed that was a typical neutron star lmxb @xcite . the source has been detected once in hard x - rays ( 35150 kev ) by sigma , making one of several neutron star lmxbs with states similar to the low - hard states of black hole binaries , but during several other observations no hard tail was detected @xcite . recently , attention has focused on because it is one of only four sources displaying both coherent oscillations during a type i x - ray burst ( at 524 hz ; @xcite ) and twin kilohertz quasi - periodic oscillations with an approximately constant separation in fourier frequency ; in the measured frequency difference is close to half of the burst frequency @xcite . in this paper we report on a comprehensive analysis of the nine x - ray bursts that were observed during observations of with the _ rossi x - ray timing explorer _ ( _ rxte _ ) between 1996 august and 1999 february . the bursts occurred while the persistent luminosity varied systematically , which allows us to study how the properties of the bursts correlate with the spectral states of the source . five of the bursts exhibited coherent oscillations , and we systematically investigate what conditions correspond to the presence of these oscillations . finally , we have tracked the evolution of the phases of the burst oscillations , which allows us to place tight constraints on any long - term change in oscillation frequency , to derive upper limits on the strength of sub - harmonics and harmonics of the observed burst frequency , and to determine the strength and energy spectrum of the pulsations . to place this work in context , we begin by reviewing briefly the phenomenology and theory of type i x - ray bursts . in section 1.1 , we examine how the properties of bursts have been observed to correlate with the properties of the persistent emission in several sources , and how theories address these correlations . in section 1.2 , we discuss the observations of nearly coherent oscillations during type i x - ray bursts and their possible connection with the spin period of the underlying neutron star . the remainder of the paper concerns the analysis ( section 2 ) and interpretation ( section 3 ) of _ rxte_observations of . type i x - ray bursts have been observed from more than 45 low - mass x - ray binary systems ( lmxbs ; see @xcite for a review , and @xcite for some recent discoveries ) . these bursts occur when unstable nuclear burning of helium or hydrogen ignites matter accreted onto the surface of the neutron star primary . the models reproduce the observed rise times ( s ) , durations ( minute ) , and recurrence times ( ) of bursts , spectral softening in the burst tails , and total energies of the bursts ( @xmath2 ergs ) . detailed models predict variations in the frequency and strength of bursts from a single source due to changes in the composition of the burning material , which is in turn determined by the metallicity of the matter accreted onto the neutron star , the amount of hydrogen burned during the time between bursts , and the amount of fuel left - over from the previous burst @xcite . in addition , variations from source to source are expected because of variations in the core temperatures of the neutron stars and accretion rate @xcite . the theoretical models that are most readily compared to observations of burst sources are those that predict how burst properties in an individual system should change as the accretion rate onto the neutron star varies . of the forty - odd bursting lmxbs , only about ten have had their burst and persistent emission properties correlated ; for the majority of sources , either few bursts are seen , or several bursts are seen in a single luminosity state @xcite . nonetheless , studies of these ten sources reveal some intriguing relationships between the persistent emission and the recurrence times , durations , and energetics of the bursts which can be compared to theoretical models . the model of @xcite predicts that x - ray bursts should occur in three regimes depending on the accretion rate ( @xmath3 ) , which they label cases 1 , 2 , and 3 . we denote these accretion rates in units of the eddington accretion rate , @xmath4 . at the lowest accretion rates ( @xmath5@xmath4 , case 3 ) , the temperature in the burning layer is too low for stable hydrogen burning ; the hydrogen ignites unstably , in turn triggering helium burning , which produces a type i x - ray burst in a hydrogen - rich environment . at higher accretion rates ( @xmath6@xmath4 , case 2 ) , hydrogen burns stably via the hot cno cycle and is converted to helium as quickly as material is accreted onto the neutron star . a pure helium layer develops at the base of the accreted material , and heats steadily until a pure helium burst is triggered . at higher accretion rates ( @xmath7@xmath4 , case 1 ) , material is accreted faster than it can be consumed by hydrogen burning ( which is limited by the rate of @xmath8-decays in the cno cycle ) , so that the helium ignites unstably in the h - rich environment ( see bildsten 1998 for dependences of these accretion rates on the metallicities ) . at higher accretion rates ( @xmath9@xmath4 ) , stable helium burning becomes important on the surface of the neutron star , which depletes the primary fuel reserves and causes bursts to occur less frequently . @xcite point out that the accretion rate inferred from the luminosity of any given neutron star lmxb that exhibits bursts can range from @xmath10 @xmath4 , so that an individual source that varies in luminosity should exhibit changes in its bursting behavior between case 2 and case 1 . this model provides three predictions . first , as the accretion rate increases , the column of material above the burning layer builds more quickly , and thus the time required to reach the critical temperature for unstable helium burning decreases . this is consistent with the increase in burst rate with persistent flux observed in @xcite and @xcite . at the highest accretion rates , steady helium burning reduces the amount of fuel for x - ray bursts , which causes bursts to occur less often , as seen in exo 0748@xmath11676 @xcite and @xcite , or not at all , as observed in @xcite and gx 3@xmath121 @xcite . however , no correlation was found between persistent flux and burst recurrence times in ser x-1 @xcite , @xcite , and @xcite , which suggests either that an additional mechanism may control how frequently bursts occur , or that the persistent flux is not a good measure of the accretion rate in these sources ( see below ) . second , the ratio of the energy released in the bursts to that released by accretion between bursts should be larger when large amounts of hydrogen are burned during a burst , and should decrease as more helium is added to the nuclear burning layer . assuming that all of the accreted material is eventually consumed in a burst , this ratio is simply the amount of energy released by the nuclear reactions the cno cycle for hydrogen burning , and the triple-@xmath13 process for helium burning divided by the gravitational energy released by the accreted material . historically , the inverse of this ratio is used , and defined as @xmath13 . the expected values for @xmath13 are @xmath14 , where the lower value corresponds to pure hydrogen burning , which releases about four times the energy per nucleon than helium burning does . the value of @xmath13 was observed to increase as expected ( from 12 to 75 ) as the persistent flux from exo 0748@xmath11763 increased by a factor of five @xcite . however , several sources exhibit @xmath13 values as large as @xmath15 , which suggests that both hydrogen and helium burn steadily during the periods of steady accretion between bursts @xcite . third , the larger the fraction of helium in a burst , the faster nuclear energy will be released ( via the triple - alpha process ) during the helium flash , so that the peak flux of the burst will be higher @xcite . conversely , if more hydrogen is present in the burning material , as in cases 1 and 3 , then helium will be diluted , slowing the rise of the burst . the hydrogen will , however , burn slowly in the tail of the burst via proton capture onto the products of the helium burning . thus the fastest , most intense bursts should occur under case 2 , which corresponds to low accretion rates . surprisingly , this is precisely contrary to the sense of the majority of the observations . the decay time scales of bursts from @xcite , @xcite , and @xcite have all been reported to decrease with increasing flux . the peak fluxes of bursts were found to increase as the persistent flux increased in exo 0748@xmath11676 @xcite , @xcite , and ser x-1 @xcite , although @xcite and @xcite exhibit no apparent correlation between the peak fluxes of bursts and the persistent flux . a survey of ten lmxbs by @xcite reveals a global decrease in burst duration with increasing flux . the sense of these observations is that the fastest , most intense bursts occur at high accretion rates . the origin of this discrepancy is uncertain ( see bildsten 2000 for a recent discussion ) . it is apparent that for each correlation mentioned above , only a few sources can be cited as examples . this is largely because two systematic difficulties interfere with observation of these effects . first , recurrence times and @xmath13 values for bursts are difficult to measure because most satellites observe from low earth orbit , where the earth occults most sources for a significant fraction of every hour . since bursts recur on a time scale of hours , many bursts are missed during occultation . only exosat , with its 91 hour orbit , was relatively free from this difficulty . second , the persistent flux from an lmxb is not a monotonic function of accretion rate . as the trends predicted by the model of fujimoto , hanawa , & miyaji ( 1981 ) are a function of the accretion rate , it is not surprising that correlations between burst properties and persistent flux would not be evident in every source . a better measure of accretion rate is provided by the color - color diagram @xcite . indeed , in the atoll source both the burst duration and the apparent temperature of the emission in the burst tails were found to be correlated with the position on the color - color diagram , but not with the flux @xcite . observations using instruments prior to the exosat mission did not provide sufficient timing and/or spectral information to deduce the spectral state of lmxbs on a color - color diagram unambiguously . fortunately , the excellent time and energy resolution available with _ rxte _ allows us to examine the relationship between burst properties and accretion rate in sources such as . although type i x - ray bursts are useful measures of the energetics , time scales and recurrence frequencies of nuclear burning on neutron stars , the bursts are only observed after the energy has propagated through the neutron star envelope , and so they are not direct probes of the burning layer itself . the discovery of near - coherent oscillations during type i bursts from several sources using data from the proportional counter array ( pca ; @xcite of the _ rossi x - ray timing explorer _ ( _ rxte _ ; @xcite ) has opened a new and extremely valuable window onto x - ray bursts and the neutron stars underlying them ( see @xcite for a review ) . the oscillations have been seen in both rising and cooling phases of the bursts and have extremely high coherence values ( @xmath16 ) , although the pulse frequency often drifts by up to 2 hz over the course of the burst @xcite . in particular , the oscillations sometimes relax towards an asymptotic frequency which is identical in bursts separated by several years stable enough ( to less than one part in one thousand ) that it has been suggested it might be used to constrain the binary mass function @xcite . the obvious candidate for a pulsation of this high coherence and long - term stability is the spin period of the neutron star @xcite . under this interpretation , it is usually proposed that the modulation is the result of localized burst emission ( a `` hot spot '' ) on the neutron star surface . several lines of argument support this hypothesis . first , the fractional amplitudes of the oscillations in the rise of bursts are generally strongest on the leading edge of the bursts , and are undetectable at the burst peak when a large portion of the neutron star surface is in conflagration @xcite . second , the rms strength of the oscillations in the rise of the bursts is very high , up to @xmath1730% in @xcite and @xmath1750% in @xcite ; note that in both cases the authors report the half - amplitude , which is a factor of 1.4 larger than the rms amplitude ) . this would be expected if the initial burning is confined to a small area on the neutron star surface @xcite . finally although it is not understood how a hot spot could survive the initial conflagration on the neutron star pulse phase - resolved spectroscopy of an oscillation seen in the tail of a burst from indicates that the flux modulation is accompanied by a modulation of the blackbody temperature of the spectrum , again as expected ( strohmayer 1998a ) . a similar conclusion has been reached for a burst from through the detection of soft lags in the burst photons , which are attributed to doppler shifts of the light from the fast - moving hot spot @xcite . the favored model for the drifts in frequency is that they indicate an expansion of the burning layer by 1030 m at the burst start . during the expansion , the rotation of the burning layer slows , only to relax back to the neutron star surface and spin up again later in the burst @xcite . cumming & bildsten ( 2000 ) have shown that the hydrostatic expansion expected during the burst is consistent with that observed , and that the shear layers are likely stabilized by the strong stratification in the atmosphere . consistent with this hydrostatic model , @xcite observed a decrease in frequency of the burst pulsations during the tail of a burst from which was coincident with a re - heating episode . still , there are several unanswered questions regarding the hot spot interpretation . even in sources where the burst pulsations are seen , not all bursts show them , which suggests that the creation of a hot spot requires special conditions . ( until now , no systematic search for correlations between the presence of coherent oscillations and the properties of the bursts or the persistent emission has been carried out . ) moreover , as mentioned above , it is not known how a hot spot could remain on the surface of the neutron star after the initial burning phase , which presumably envelops the whole neutron star @xcite . finally , @xcite finds that in one burst from ( his burst c ) the pulsations have an `` asymptotic '' frequency that is significantly less than the highest frequency observed during the burst , which is hard to reconcile with the simple picture of a burning layer contracting and re - coupling to the underlying neutron star . the importance of the burst oscillations has been further enhanced by the detection of quasi - periodic oscillations ( qpos ) in the 2001200 hz frequency range in power density spectra ( pds ) of the persistent emission from numerous lmxbs . these so - called khz qpos ( for reviews , see @xcite , @xcite ) may reflect the relativistic motions of material at the inner edge of the accretion disk in these systems @xcite . the relative ubiquity of pairs of khz qpos with the twin peaks separated by a frequency difference that stays roughly constant as the frequencies of the peaks increase @xcite leads naturally to beat - frequency models which identify as the spin frequency of the neutron star , @xcite . initially this interpretation was supported by the observation that was consistent ( within errors ) with the burst frequency , or half that , in sources where both phenomena were observed . this coincidence was evident in and ( @xmath18 ) , and in and ( @xmath19 ) . ( note that in the burst - pulsation sources and twin khz peaks have not been detected to date . ) the detection of a sub - harmonic signal in @xcite has lent the idea further support by demonstrating that the strongest signal in the bursts from this source is actually at the first harmonic of the spin period , @xmath20 , so that @xmath21 . however , it has also since been observed that ( 1 ) is not constant in @xcite , @xcite , and @xcite ; ( 2 ) khz qpo data from all other sources are equally consistent with a changing or a constant @xcite ; and , ( 3 ) the `` equalities '' in and do not in fact hold ( @xmath22 in , @xcite ; and @xmath23 in , @xcite . one response to this dilemma is proposed by @xcite , who suggest that the decreasing difference frequency is that of a differentially rotating magnetosphere . in their model , the lower khz qpo represents the keplerian frequency at the inner disk , and the upper khz qpo occurs at an upper hybrid frequency of keplerian blobs in elliptical orbits , with the frequency determined by the coriolis force acting in the frame of reference rotating with the neutron star magnetosphere . another set of models states that the upper and lower khz qpos are the general relativistic keplerian and periastron precession frequencies near the inner edge of the accretion disk @xcite . the difference frequency in these latter models is unrelated to the spin period of the neutron star , and the apparent relationship between and is taken to be a coincidence . it is therefore critical to determine the nature of the coherent oscillations observed during bursts in order to distinguish between models for the khz qpos . provides crucial observational evidence for testing the current models , as it is one of only four sources for which both the burst frequency and the difference frequency have been measured . we have analyzed all 27 observations of taken between 1996 july 14 to 1999 february 27 with the proportional counter array aboard _ rxte _ ( table [ obs ] ) . during these observations , nine x - ray bursts are seen , including the one previously reported by @xcite . in order to characterize the persistent emission from , we have created background - subtracted light curves of the x - ray intensity of from standard2 data for each observation . we have calculated the ratios of the count rates in the 3.44.8 kev @xmath24 2.03.4 kev bands ( soft color ) , and in the 8.518.0 kev @xmath24 4.88.5 kev bands ( hard color ) for 256 s interval during each observations . the mean count rates and hardness ratios , excluding time intervals containing bursts , are listed in table [ obs ] . figure [ cc ] displays the color - color diagram for this set of observations . the states with the lowest observed count rates correspond to points with a hard color of @xmath25 , while those points with higher count rates have a hard color @xmath26 . the count rate from the source is not a monotonic function of position on this diagram ( table [ obs ] ) . although motion on the branch with lower hard color is observed in the course of several hours , the transition from a hard color @xmath25 to @xmath26 occurred between observations separated by more than a month . we have produced power - density spectra ( pds ) of photons in the 260 kev _ rxte _ band for each observation , in order to determine whether exhibits the canonical branches of the color - color diagram for lmxbs @xcite . the pds of each observation have been created by combining averaged pds from data with @xmath27 s time resolution in 64 s intervals with a single pds of data with @xmath28 s time resolution for the entire observation . bursts are excluded from the pds . we have found that pds are of two types , as represented in figure [ pds ] . figure [ pds]a is representative of pds from observations with hard colors @xmath29 , which can be described as the sum of relatively weak ( 2% rms ) low frequency noise below 0.1 hz and strong ( 13% rms ) flat - topped high frequency noise that decreases in power above a few hz . figure [ pds]b is representative of pds from observations with hard colors @xmath30 , and which can be characterized as a combination of @xmath31% rms low frequency noise below 1 hz and weak power ( 2% rms ) at high frequencies . strong ( @xmath32% rms ) qpos are not evident below 100 hz in the pds from either state , and although single khz qpos are evident on a few occasions in photons with energy above 5 kev ( not shown ) , twin khz qpos are only evident in the observation reported on by wijnands & van der klis ( 1997 ) . taken together , the color - color diagram and pds indicate that is indeed an atoll source , as suggested by @xcite ( for a review of z and atoll lmxbs , see @xcite ; see @xcite for a similar color - color diagram of the atoll source as observed with the pca on _ rxte _ ) . the state with hard color @xmath25 is the island state , which is thought to represent a low accretion rate ( @xmath17 0.01 times the eddington rate , @xmath33 ) , while the state with hard color @xmath26 is the banana state , thought to occur at a somewhat higher accretion rate ( @xmath34 ) . this classification of is consistent with the presence of strong x - ray bursts only two z - sources , gx 17@xmath122 @xcite and cyg x@xmath112 @xcite , exhibit type i x - ray bursts . an arrow has been drawn on the color - color diagram to indicate the direction believed to correspond to increasing accretion rate . we have estimated the source flux during each observation ( table [ obs ] ) by fitting the standard2 data summed over each _ orbit with a two - component model , consisting of an exponentially cut - off power - law ( @xmath35 , where @xmath36 is the photon energy , @xmath37 is the power law photon index , @xmath38 is the cut - off energy , and @xmath39 is the photon flux at 1 kev ) and a gaussian line at 6.4 kev , which is likely due to iron emission . we have accounted for interstellar absorption with a low - energy cut - off equivalent to a column density of 10@xmath40 @xmath41 of hydrogen ( according to the value derived with _ rosat _ , @xcite . the time intervals during bursts have been excluded from the spectra . although we fit the spectra between 2.525 kev , the flux was calculated from the model between 218 kev , to correspond with the energy range used for the color - color diagram . the reduced chi - squared value for this model is less than 1.1 for all observations except , , and , for which the reduced chi - squared values are near two . the derived flux is not significantly affected by the poor values of reduced - chi squared , and we report all values in table [ obs ] . for a more detailed discussion of the x - ray spectrum of as observed by _ rxte _ see @xcite . table [ burst ] lists the times ( corrected to the earth - sun barycenter ) of the nine bursts observed with the pca . we have integrated spectra for each 0.25 s interval during these bursts using combinations of the binned and event mode data for each observation . this provides 32 energy channels for the first burst and 74 energy channels for the remainder of the bursts . since the effect of the x - ray burst on the spectrum of the persistent emission is unknown , it is not clear what `` background '' should be subtracted from the total received flux in order to obtain the true burst spectrum . therefore we have estimated the background in two independent ways , representing two extremes of approach : ( 1 ) we have used the ftool pcabackest to determine the instrumental and cosmic background , including both the persistent and burst flux in our fits ; and ( 2 ) we have used the spectrum of 100 s of data prior to the burst as background , subtracting the persistent emission under the assumption that it is unaffected by the burst . ( we note that it is possible that persistent emission increases during the burst , with accretion induced by the burst itself , but we have no theoretical guidelines for modeling such a process and therefore ignore it . ) we have fit the resulting background - subtracted spectra between 2.525 kev with a single blackbody ( without absorption , since fits with absorption favor very small values of the column density and do not improve chi - squared significantly ) , which provides two parameters : the apparent temperature and solid angle of the emission region . from the solid angle we can derive the apparent radius of the emission region given a derived or assumed distance to the source . from the model parameters , we calculate the bolometric flux for each interval during the burst , @xmath42 . we have compared spectral fits after subtracting the background in these two ways , and find that the resulting parameters are very similar , although using the pcabackest estimate of the background increases @xmath43 for the fit by a factor of 1.31.6 relative to fits with the pre - burst emission subtracted . the increase in @xmath43 is almost certainly due to the presence of flux from steady accretion in the spectra that remain after subtracting the pcabackest estimate . to keep the spectral model simple , _ we report results only for spectra obtained after subtracting the pre - burst emission . _ the resulting spectral parameters should be interpreted with some caution , because it is not likely that the burst emission is a pure black body : on the contrary , electron scattering in the neutron star atmosphere is likely to cause systematic deviations in the spectrum which would result in systematically high ( and thus low ) at a given flux ( see also the discussion in section 3 ) . we have analyzed the profile and strength of the bursts by finding the rise time @xmath44 , start time of the decay @xmath45 , decay time scale @xmath46 , maximum flux @xmath47 , fluence @xmath48 , and duration @xmath49 for each burst ( table [ burst ] ) . the rise time @xmath44 has been defined as the time for the burst to increase from 25% to 90% of the peak flux . the start time of the decay @xmath45 is the approximate time when the flux begins decreasing exponentially , measured from the start of the burst . we have fit the decay in flux of each of these bursts with either one or two successive exponential functions , @xmath50 , ( for four and five bursts , respectively , depending on the value of reduced chi - squared for a fit with a single exponential ) in order to estimate the time - scale for the decay of the bursts . the exponential fits work well ; for all of the bursts , @xmath51 . finally , we have summed the flux from the burst up to the start of the burst decay with the flux in the burst tail , as determined by integrating the exponential fit , to determine the fluence of the burst @xmath48 . the spectral fits reveal which of the bursts exhibited an episode of radius expansion ( last column in table [ burst ] ) . we have found that four bursts exhibit radius expansion episodes . in all cases the effective area of the emitting region expands by a factor of 2.55.2 . since the maximum luminosity during the radius expansion episodes is thought to represent the eddington luminosity of the neutron star @xmath52 , we can use the maximum peak flux from all of the radius - expansion bursts , @xmath53 erg @xmath41 s@xmath54 , to estimate the distance to the source @xmath55 via @xmath56 erg s@xmath54(see van paradijs , pennix , & lewin , 1988 ) , where we have taken the @xmath52 corresponding to a pure helium photosphere for a 1.4 neutron star , and the redshift of the photosphere to be zero when the peak flux is measured . we use the largest value because some variation in the peak flux in radius expansion bursts is to be expected , as the observed eddington luminosity decreases when a lower fraction of helium is present in the photosphere , when there is any anisotropy in the burst emission , and when the redshift of the photosphere at its maximum observed expansion is larger . our estimate places an upper limit to the distance to of @xmath57 kpc . the previous estimate of the distance to was derived from the flux from burst 1 in the same manner as above by @xcite . the interstellar absorption column may be variable by as much as a factor of two , and therefore is not a useful measure of the distance to the source ( see @xcite . the spectral parameters from three representative bursts a large burst without radius expansion ( burst 3 ) , a large burst with radius expansion ( burst 9 ) , and a weak burst without radius expansion ( of which we observe only one case , burst 7 ) are displayed in figure [ bspec ] . the apparent radius of the blackbody has been plotted assuming that the distance to is 7 kpc . the radius expansion bursts tend to be shorter than the bursts without radius expansion ( as indicated by @xmath58 in table [ burst ] ) . we have indicated the spectral state during which the bursts occurred by the numbers in figure [ cc ] . the placement of the numbers indicates the colors of the persistent emission immediately before the burst , and a square has been drawn around the burst number if the burst exhibited radius expansion . all of the bursts which exhibit radius expansion episodes ( bursts 1 , 2 , 8 , and 9 ) occurred while was on the banana branch , with hard color less than 0.5 , while all of the large bursts without radius expansion episodes ( bursts 3 , 4 , 5 , and 6 ) occurred while was in the island state , with hard color greater than 0.6 . the single small burst 7 occurred on the banana branch , which is not surprising given its small @xmath58 and short decay time scale . this behavior was also seen in exosat observations of other sources ( e.g. @xcite ) . following gottwald ( 1986 ) , in figure [ lvsr ] we have plotted the apparent radius during bursts from ( top panels ) and the bolometric flux in the burst ( bottom panels ) as a function of apparent temperature . the numbers in the figure indicate the burst from which the data have been taken . we have only plotted points where the burst flux is greater than 15% of the peak flux from the burst , because at lower fluxes the uncertainty on the spectral parameters becomes large ( see figure [ bspec ] ) . the panels on the left , for values of the persistent flux greater than @xmath59 erg @xmath41 s@xmath54 , represent the bursts which occurred on the banana branch . the panels on the right , for values of the persistent flux less than @xmath59 erg @xmath41 s@xmath54 , represent the bursts which occurred in the island state . a two - branch pattern in the @xmath60 vs. @xmath61 relationship is evident in bursts from the banana branch ( top - left panel of figure [ lvsr ] ) , which is not present for the bursts from the island state ( top - right panel ) . the right - side ( @xmath62 kev ) excursion to high @xmath63 km represents the radius expansion phase at the start of the burst ( note that burst 7 exhibits no radius expansion , and therefore does not follow this track ) , while the excursion at lower temperatures ( @xmath64 kev , @xmath65 km ) represents an increase in the apparent emitting area in the tails of the fast bursts ( e.g.fig . [ bspec]b ) . it is impossible to precisely determine the time interval between bursts and the ratio of energy released in the persistent to the bursting luminosity ( @xmath13 ) , because was occulted by the earth for about a third of every 90 minute _ rxte _ orbit , and the observed bursts occurred infrequently . we can only roughly estimate the rate with which bursts occurred and the values of @xmath13 during four intervals with more than 10 ks of exposure time over the course of a few days ( see table [ obs ] ) . the values that follow are uncertain by about a factor of two because of the small number of bursts detected . for 1997 october 28 to 29 , the source was on the banana branch and one burst was observed during 13.9 hours of exposure , suggesting an approximate rate of 0.07 hr@xmath54 . integrating both the persistent ( table [ obs ] ) and bursting flux ( table [ burst ] ) over the exposure time during this interval , we find @xmath66 . no bursts were observed during 13.9 hours of observations during the interval from 1998 july 31 to august 1 , also on the banana branch , which is consistent with the single burst seen in the first interval . was in the island state for 1998 october 2 to 6 , and the approximate burst rate was @xmath67 hr@xmath54 with @xmath68 . during 1999 february 22 to 27 the source was again on the banana branch , and the burst rate was @xmath69 hr@xmath54 with @xmath70 . the burst rate does not seem to show any correlation with the spectral state of the source , which suggests that some parameter other than the accretion rate determines the frequency with which bursts occur . on the other hand , the values of @xmath13 are at least a factor of six higher when is on the banana branch ( @xmath71 ) than when it is in the island state ( @xmath72 ) . this is consistent with the previous exosat phenomenology as well , which pointed to higher @xmath13 values when the bursts were short and radius expansion was prevalent , and thus to more helium - dominated bursts . we have attempted to determine whether bursts could have occurred regularly during the four time intervals above by assuming that they recur with times distributed in a gaussian fashion about a mean time interval . we have determined the probability that we could have observed bursts at the times in table [ burst ] given recurrence intervals between 0.2 and 20 hours . depending on the interval , we find that either ( 1 ) no constant mean recurrence time is consistent with the observed bursts ( and lack thereof ) ; or ( 2 ) the only consistent recurrence times are longer than our estimate of the burst times above , which would suggest that we were `` lucky '' to observe the bursts that we did . an alternative possibility is that bursts from occur in clusters , similar to the behavior of ser x-1 @xcite , @xcite , and @xcite . in a uniform search for burst pulsations , we have produced power density spectra ( pds ) for 1 s intervals for the first 15 seconds of the bursts ( starting from the second before the burst rise ) , using data from the 220 kev energy range binned to a time resolution of @xmath27 s. we have searched these pds over 5 hz frequency intervals centered at 262 , 524 , and 1048 hz for signals at the sub - harmonic , fundamental , and first harmonic of the signal observed by @xcite in the first burst . our detection threshold power ( 99%-confidence ) is 24.4 when normalized according to the criteria of leahy ( 1983 ) , taking into account all of the examined bins . we then took into account the distribution of noise powers in the pds , and inverted the function of @xcite using the algorithm in the appendix of @xcite to calculate either 1-@xmath73 uncertainties on measured powers , or 99% upper limits to the true powers when none was detected significantly . assuming any signal power ( or upper limit ) is produced by a sinusoidal variation of amplitude @xmath74 , we can invert the relationship @xmath75 where @xmath76 is the power in frequency bin @xmath77 , and @xmath78 is the nyquist frequency of the pds , in order to find the oscillation amplitude @xcite . table [ osc ] reports the rms amplitudes of the largest detected oscillations , or upper limits on oscillations derived from the largest observed power , for each of the bursts . five signals have been detected at a frequency of 524 hz , and no significant signals have been detected at the sub - harmonic or first harmonic . the signal in burst 7 is significant with only 90%-confidence given our broad search parameters , but since there is a ( much smaller ) 0.2% chance that the signal could be due to noise in a search restricted to the first second of the burst rise , and since the significance of these pulsations has been confirmed by phase modeling ( see below ) , we treat this signal as a detection . the rms amplitude of the oscillation in burst 7 is large only because the count rate in the rise of the burst is a factor of 2 smaller than the other bursts ( compare table [ burst ] ) . a brief comparison of table [ osc ] and figure [ cc ] reveals a striking discovery . bursts during which coherent oscillations were observed have been marked with a diamond in figure [ cc ] . along with table [ osc ] , this figure reveals that all four of the burst signals that occur after the rise of the burst were found in bursts that exhibited radius expansion episodes ( bursts 1 , 2 , 8 , and 9 ) . as a signal is also found in the rise of burst 7 , all of the bursts seen on the banana branch exhibited coherent oscillations . on the other hand , none of the bursts in the island state ( bursts 3 , 4 , 5 , and 6 ) exhibited coherent oscillations ( to the sensitivity of our search ) . for the bursts with detected oscillations , we have examined the evolution of the burst pulsations by creating a dynamic power - density spectrum ( fig . [ fev ] ) . for these plots we have executed 4 s ffts at 0.25 s intervals , oversampling the pds by a factor of two to improve frequency resolution ( see , e.g. , middleditch ( 1975 ) ; chakrabarty ( 1996 ) ) and applying a welch window function ( press 1992 ) to reduce the side - lobes on strong powers . for all bursts we have found that the strengths of the oscillations were enhanced by restricting our analysis to higher energy photons : for burst 1 we select photons with energies above 3.6 kev , and for the remaining bursts we select photons with energies above 5.1 kev . figure [ fev ] shows the character of the frequency evolution in each case . in most cases contour levels represent powers that are powers of 10@xmath54 in single - trial probability , starting with a 10% chance detection ( where the signal is likely to be noise ) at the outermost contour , and probablities of @xmath79 , @xmath80 , etc . for inner contours ; in burst 9 , however , the pulsations are significantly stronger : the first contour is at a single - trial probability of @xmath81 , with further contours spaced by powers of 10@xmath82 . for comparison purposes we also have provided , in the plots of fig . [ fev ] , the 260 kev light curve of the bursts from standard1 data ( running from 0 to 24,000 pca c s@xmath54 , full scale ) and the burst blackbody radius as derived from our spectral fits ( see above ; a distance of 7 kpc is assumed ) . in bursts 1 , 2 , 8 , and 9 the burst pulsations commenced when the blackbody radius dropped below km ; in bursts 7 and 8 a short set of oscillations occurred in the rise of the burst . finally , we have plotted the frequency evolution as derived from our phase models ( see below ; approximate plus or minus one - sigma ranges are indicated ) , which confirms in more precise terms the overall picture given by the dynamic pds . we have observed frequency evolution in all cases , except during the rise of bursts 7 and 8 , when the duration of the signal is short , @xmath831 s. the form of the frequency evolution in bursts 1 and 9 is very similar : in both cases the frequency increases by approximately 0.5 hz during the course of the burst decay , relaxing upwards in a quasi - exponential fashion . this behavior was not observed by @xcite because they created a single pds every second , providing them with a frequency resolution of only 1 hz . the frequency evolution we see in these bursts is strongly reminiscent of that seen in the burst sources @xcite and @xcite , although the magnitude of the fractional frequency shift is a factor of @xmath84 smaller . burst 2 shows the most complicated behavior , starting at a frequency of approximately 524 hz and evolving towards 525 hz in a sort of s - curve . burst 8 , in addition to showing pulsations on the burst rise , shows a substantial frequency decrease during the burst peak , joining the burst from analyzed by @xcite and the additional burst from analyzed by @xcite as counterexamples to the spin - up behavior that initially seemed to characterize the burst oscillations . in order to make a quantitative study of the form of the frequency evolution , we have performed _ phase connection _ of the burst pulsations . the technique of phase connection is familiar from traditional radio ( manchester & taylor 1977 ) and x - ray pulsar studies ( e.g. @xcite , @xcite ) , and allows for a precise characterization of the pulse train from a source even when individual pulses are not detectable in the data . our phase connection procedure works as follows . beginning with the dynamic pds , we identify the interval of strongest pulsations and their corresponding average frequency . we take each half - second of data from the burst , fold at the average frequency , test for the presence of significant modulation , and if sinusoidal modulation is present ( with @xmath990% confidence ) , cross - correlate the pulse profile with a sinusoid to determine the reference phase of the pulsation at the start of that half - second interval . we use twelve phase bins for our folded profiles ; our results are not sensitive to this choice . the uncertainty in the reference phase is determined by monte carlo ( mc ) simulation : we randomly perturb the count rate for each phase bin within the range expected from poisson counting noise and recalculate the reference phase 100 times for each phase point ; our phase uncertainty is the standard deviation of these mc phases . with reference phases and uncertainties in hand we connect the phases with a best - fit polynomial phase model . the accuracy of this model is then confirmed by epoch folding and phase referencing ( i.e. , the procedure described above ) with respect to the model : a successful model , by accurately tracking the pulsations through their full evolution , will produce phase residuals scattered around zero , whereas an inaccurate model ( where it evinces any modulation at all ) will produce phase residuals evenly covering the ( @xmath110.5 , 0.5 ) interval past the point where the model breaks down . we take reference phases to lie between @xmath110.5 and 0.5 . since the derived phase uncertainties are @xmath830.10 , chances that a -point phase model will serendipitously be `` successful '' by this criterion are entirely negligible . in some cases , the pulsations are strong enough that we can work with 0.25 s time intervals , and for the sake of better model constraints we do this for the pulsations during the rise of bursts 7 and 8 . the above procedure can be iterated , as necessary , or applied to arbitrary phase models until a satisfactory chi - squared value is obtained . in our analysis we typically added terms until the model had the flexibility to accomodate the evolution seen in the dynamic pds and , moreover , had a reasonable chi - squared value . naturally , a ( more or less ) smooth polynomial evolution of the frequency is not guaranteed a priori ; on the contrary , @xcite and @xcite have shown that piecewise - linear frequency drift models with occasional sudden jumps in @xmath85 describe the frequency evolution in several bursts from quite well . nevertheless we have found polynomial models adequate for all of the bursts . a detailed description of our phase models is presented in table [ model ] . in order to determine the approximate one - sigma ranges shown in the plots of figure [ fev ] , we have set each parameter of the fit ( in turn ) to its one - sigma limit as shown in the table , minimized @xmath86 with respect to the remaining parameters , and calculated the frequency evolution throughout the pulsation interval , as derived from the fit . the full range in frequency sampled by each point in time , in the course of sampling the different fits , was saved and determined the confidence regions that we have plotted . although this procedure falls short of the full exploration of @xmath86 space that would be required for exact limits , we feel that it captures the essence of the flexibility allowed with respect to the best - fit model . the flaws with our approximate method are most apparent in the plot for burst 2 , figure [ fev](b ) , where the uncertainty in the frequency at intermediate times during the evolution is certainly greater than shown . inspired by the form of the frequency evolution shown in the dynamic pds of bursts 1 and 9 ( fig . [ fev ] ) , and by the work of @xcite , we have also fit exponential - relaxation phase models to the data from these two bursts . note that the final polynomial phase models for these bursts have four terms each ( quadratic frequency evolution ) so that we could pursue exponential models without sacrificing any degrees of freedom . we have found that the exponential - relaxation models , @xmath87 , provide fits as good as the polynomial models in both cases , with chi - squared values of 10.8 for 8 degrees of freedom ( burst 1 ) and 13.2 for 11 degrees of freedom ( burst 9 ) , respectively . full details of the exponential - relaxation fits are shown in table [ expmodel ] . the best fit values for the asymptotic frequencies of the burst pulsations are 524.61@xmath88 and 524.48@xmath89 hz for bursts 1 and 9 respectively . the coincidence of these two values ( which are consistent at the 1.5@xmath73 level ) is remarkable , with the difference being equivalent to an orbital doppler shift ( @xmath90 ) of only 75 km s@xmath54 . this near - equality of asymptotic frequencies found using an exponential frequency evolution model has been observed in , , and as well @xcite , which suggests that the model is physically meaningful . proceeding to a joint investigation of the two physically meaningful parameters of the exponential model , and , we present in figure [ twodreg ] the two - dimensional confidence region for and , as constrained by our exponential - relaxation model fits for bursts 1 and 9 . note that values of the two other parameters , @xmath91 and ( the phase of the pulsations at @xmath92 ) , are sensitive to the exact start time of our phase model and hence depend on our instrument sensitivity . examination of figure [ twodreg ] reveals , first of all , that the fit parameters for bursts 1 and 9 are consistent at better than 1@xmath73 , as the 1@xmath73 confidence regions for the two bursts overlap near @xmath93 hz , @xmath94 s. second , the figure shows that a - degeneracy weakens our constraints at large ( , ) . the sense of this degeneracy is that increasing both and moves the final approach to a constant frequency out to later times , and so will be allowed to the extent that the frequency is still increasing when the oscillations disappear . since the pulsations in burst 1 are weaker than those in burst 9 , especially at late times , the problem is more acute for this burst . nevertheless , in examining the plot we may draw some conclusions . first , if the values for the two bursts are in fact equal ( or nearly equal ) , then the values for the bursts must also be nearly equal , within 10% . alternatively , if the values for the two bursts are in fact different , whether because of an intrinsic change in the pulse frequency or because of orbital doppler effects , then the values for the bursts are also likely to be different . the phase models in table [ model ] demonstrate full coherence of the burst pulsations from over the duration of the intervals we model . the longest of these intervals are : for burst 1 , 9.75 s or 5111 cycles ; for burst 2 , 8 s or 4194 cycles ; and for burst 9 , 9.0 s or 4718 cycles . with successful phase models for the burst pulsations we may address a number of issues . first , we can make sensitive tests for the presence of harmonics and sub - harmonics of the burst pulsations . we searched first for significant modulation at twice the main burst frequency . by multiplying the phases derived from our phase model by two , and integrating through the burst pulsation intervals indicated in table [ model ] , we have found evidence at the 2.1@xmath73 level for a modulation at the first harmonic of the 524 hz pulsations . the average power of this modulation , over all the intervals we have modeled , is 1.9@xmath95% that of the main signal . we then searched for and found no evidence for modulation at the second and third harmonics of the main signal ; 2@xmath73 upper limits are 2.3% and 1.7% the power of the main signal , respectively . we then searched for a sub - harmonic of the main signal by multiplying our derived phases by 0.5 ; as @xcite points out , we must also account for the possibility of half - phase offsets at the points of connection between different burst intervals . since the signals from bursts 1 and 9 are significantly stronger than the signals in the other bursts , it is possible that the increased number of trials required to connect all of the intervals could offset the increased sensitivity provided by coherent phase connection of those intervals . therefore we have performed this search in two stages . in the first stage , we tested for a sub - harmonic in the coherent connection of bursts 1 and 9 only ( two trials ) : we did not detect significant ( @xmath995%-confidence ) modulation in either connection , and can place a 2@xmath73 upper limit on the strength of a sub - harmonic in these bursts , relative to the main signal , of 5.5% . note that the sub - harmonic detected by @xcite in the burst - rise pulsations from was on average 19% the power of the main signal . in the second stage of the sub - harmonic search we connected all six of the burst pulsation intervals together , accounting for all possible combinations of half - phase offsets ( 32 trials ) : we did not detect significant modulation in any connection , and place a two - sigma upper limit on the average power of a sub - harmonic across all these intervals of 3.9% that of the main signal . at the same time we tested for the presence of modulation at 1.5 times the main burst frequency , and can place a two - sigma upper limit on the average power at that frequency of 4.4% that of the main signal . the last two points of the phase model for burst 2 may be used to derive a `` peak frequency '' for the pulsations in this burst of [email protected] hz . although this frequency is not as well constrained as the asymptotic frequencies of bursts 1 and 9 , it is higher than the burst 9 at a 3.2@xmath73 level of confidence , and requires a relative doppler shift of 340@xmath96100 km s@xmath54 to explain , if the we measure in fact corresponds to the spin frequency of the neutron star since the burning layer may expand to produce lower frequencies , but can not `` contract '' to produce higher ones . for the rest of our analysis we concentrate on the pulsations from bursts 1 and 9 , which have the highest signal - to - noise . we have folded data with @xmath27 s time resolution in five energy channels ( starting at 3 , 5 , 7 , 10 , and 13 kev ) from bursts 1 and 9 about the polynomial phase models . fitting a sinusoid to the folded profile in each of five energy bands , we measure the strength of the oscillation and the phase offset in each band . the reduced chi - squared values for the sinusoidal fits were less than 1.3 for all energy bands . we have found that the rms amplitude of the burst oscillation increases monotonically with energy for both bursts ; the sense of this increase is shown in figure [ oscvse ] . the measured phase of the oscillations in different energy bands provides marginal evidence for a soft phase lag in the 79 kev and 1013 kev bands relative to the 35 kev band in burst 1 ( @[email protected] and @[email protected] radians , respectively ) , and none in burst 9 . computing the fourier phase lag for a signal at 524 hz ( with 1 hz frequency resolution ) without adjusting for the polynomial phase model , we have found a marginally significant hard lag of [email protected] radians between the 57 kev and 725 kev bands in burst 9 , and none in burst 1 . given these ambiguous detections , we do not feel our measurements are significant evidence of a phase lag in the pulsed signal . finally , we have been able to accurately characterize the evolution of the strength of the pulsations over the course of the bursts ( fig . [ expev ] ) . without an accurate phase model , the strength of the pulsations would otherwise be very difficult to measure when they first appear and their frequency is changing the fastest . it appears that the strength of the burst pulsations ( relative to the burst flux ) in burst 1 are consistent with a 7% rms modulation for most of the burst , rising to a maximum of 10% before disappearing at the end ; in burst 9 , the data show a significant increase at the start of the pulsations , and are then consistent with a more or less constant 8% rms modulation for the duration of the pulsations . in both bursts , the pulsations are initially detected at a lower level of 3% rms . we have calculated the strength of the oscillations in the other bursts where they are detected , although we do not present the results as figures : the pulsations in burst 2 increase from @xmath97% rms to @xmath98% rms over the course of the interval when pulsations are detected ; those in burst 7 are consistent with @xmath99% rms throughout the 0.75 s of detected pulsations ; those in burst 8 have @xmath100% rms at start , decreasing to @xmath101% before disappearing . figure [ expev ] also shows our phase model fits and residuals for bursts 1 and 9 , presented in the form of a `` phase loss '' plot with the running phase compared to that for a constant - frequency pulsation at the fitted asymptotic frequency . according to our models , the burst pulsations lose 3.9@xmath102 cycles compared to the constant in burst 1 , and 2.8@xmath103 cycles compared to the ( different ) constant in burst 9 . note that due to the - degeneracy in the fits , these values are not well constrained from above ( especially for burst 1 ) : our two - sigma upper limits on the total phase loss are 19.4 and 4.3 cycles for bursts 1 and 9 , respectively . our results allow us to address a number of important aspects of the phenomenology of type i x - ray bursts and their burst pulsations . the nine bursts that we have observed from divide into two populations , as shown in table [ burst ] and figure [ cc ] . when the source is in the island state and the persistent flux is low , bursts have relatively broad maxima ( as measured by the time until the start of the burst decay ) , long decay time scales ( @xmath58 ) , large fluences , and no evidence for radius expansion ( bursts 3 , 4 , 5 , and 6 ) ; following @xcite we call these _ slow bursts . _ when the source is on the banana branch and the persistent flux is high , the peaks of bursts are sharper , the bursts are shorter , radius expansion episodes are present , and burst fluences are smaller ( bursts 1 , 2 , 8 and 9 ) ; we call these _ fast bursts . _ burst 7 is particularly weak , but still exhibits the short decay , narrow maximum , and small @xmath58 characteristic of fast bursts . when we compare the ratio of the energy released in the persistent luminosity to that in type i x - ray bursts , @xmath13 , we find yet another distinction : the slow bursts which occur in the island state correspond to an @xmath72 , while the fast bursts which occur on the banana branch correspond to higher @xmath104 . the decrease in burst duration with increasing accretion rate is consistent with the global relationship which @xcite reported for their sample of 10 lmxbs . note however that they found burst properties to span a continuum between the slow and fast populations in a sample of many burst sources , and that itself may display a continuum of bursting properties should the source be observed for a longer period of time at an intermediate value of the accretion rate as inferred from the color - color diagram ( figure [ cc ] ) . the work of hanawa , fujimoto , & miyaji ( 1981 ) ( see also @xcite , @xcite ) suggests that the fast bursts ( and values of @xmath105 ) result from the burning of helium - rich material . the association between fast bursts and radius expansion has previously been noted by several authors ( see , e.g. , @xcite for exo 0748@xmath11676 and @xcite for ) , and indicates that the nuclear energy of these bursts is released so rapidly that the eddington luminosity is exceeded , causing substantial expansion of the neutron star s photosphere . hanawa , fujimoto , & miyaji ( 1981 ) predict that helium - rich bursts should occur when the local accretion rate in the area responsible for the burst is low ( case 2 of section 1 , above ) . although this may seem to be contradicted by our finding that fast bursts occur at relatively high accretion rates , the paradox can be resolved if accretion takes place over a larger fraction of the neutron star when the total accretion rate is high , such that the accretion rate per unit area is reduced ( bildsten 2000 ) . evidence for this hypothesis may be found in spectral fits to the burst emission . @xcite noted that the fitted blackbody radii during the tails of fast bursts from exo 0748@xmath11676 were systematically larger than the radii for the slow bursts and bildsten ( 2000 ) showed that this is adequate to change the local accretion rate . similarly , we find that the apparent radii in the tails of fast bursts from are systematically larger than those of slow bursts ( figure [ lvsr ] ) ; this behavior can be seen in the spectral evolution of bursts 3 and 9 in figure [ bspec ] . if this difference in apparent blackbody radii reflects a physical increase in the area of the burst emission , that might explain why the properties of bursts observed at high total accretion rates are those expected for bursts occurring at a low local accretion rate . the behavior of burst 7 runs somewhat counter to this trend : although this burst has significantly less fluence than the other fast bursts , it exhibits a larger increase in blackbody radius during the tail of the burst . if a burst only ignites when sufficient material per unit area is accreted to produce the critical density and temperature , then larger burst areas should correspond reliably to larger burst fluences . on the other hand , if a burst occurred during the earth occultation 2575 minutes prior to burst 7 , then the observed burst would be similar to other weak bursts with short waiting times and relatively large apparent radii from @xcite and exo 0748@xmath11676 @xcite . these weak bursts may be triggered by unburnt nuclear fuel from the previous burst @xcite . it is also possible that there is a systematic difference in the neutron star atmosphere or the accretion flow which causes the spectrum of bursts on the banana branch to systematically deviate from a blackbody due to electron scattering . london , taam & howard ( 1984 , 1986 ) demonstrated that at luminosities close to the eddington value , the ratio @xmath106 of the apparent temperature derived from a blackbody fit to the effective temperature of the emission region increases as increases . since @xmath107 , and for a constant physical emission area @xmath108 , we have @xmath109 . thus , as the burst cools , and @xmath110 decrease together , and will appear to increase ( see also @xcite ) . if the anti - correlation between and in figure [ lvsr ] is indeed due to electron scattering , then the trend is not a probe of accretion onto the neutron star surface , but indicates that the atmosphere or immediate environment of the neutron star is fundamentally different in the banana and island states . however , the results of london were originally considered important in explaining apparent super - eddington temperatures in the peaks of radius - expansion bursts , and it is not clear whether the theory applies to the cooling tails of bursts , when the temperature is nearly a factor of two smaller . london , taam , & howard ( 1984 , 1986 ) find that the opposite correlation between @xmath110 and holds at low . it has previously been observed that coherent oscillations are not present in every burst from and @xcite , although no systematic study has been previously carried out to determine whether the properties of the bursts or persistent emission from those sources are correlated with the presence of detectable coherent oscillations . we find that in this differentiation is strictly consistent with the two burst populations we identify : all of the fast bursts exhibit near - coherent oscillations and none of the slow bursts do ( to limits of % rms in any single second ; see table [ burst ] ) . however , it is worth differentiating the oscillations found in the rise of the bursts from those occurring in the burst tails , as it is more difficult to understand how the latter form . only two instances of coherent oscillations in the burst rise are observed , so the question remains as to why some of the fast bursts exhibit burst - rise pulsations and others do not . oscillations in the rise of bursts are not tied to radius expansion episodes , as the oscillations are observed in burst 7 from and in several bursts without radius expansion in @xcite . on the other hand , all of the oscillations which are observed in the tails of the bursts appear immediately after the contraction phase of a radius expansion episode ( as first noted by @xcite ; no counter - example of coherent oscillations in the tail of a burst without radius expansion has been reported among the other sources to our knowledge ) . this suggests that the radius expansion episode leaves behind an anisotropy in the neutron star photosphere that resembles a rotating hot spot . in any case , the general association between fast bursts and coherent oscillations suggests that the process which produces the oscillations is related to the particularly intense helium - rich burning in these bursts . our analysis of the burst pulsations raises some intriguing questions about the spin - frequency interpretation of the burst pulsations in . the best - fit values of the asymptotic burst pulse frequency for bursts 1 and 9 are 524.61@xmath88 and 524.48@xmath89 hz respectively , are consistent at the 1.5@xmath73 level . the 0.13 hz frequency difference could be produced by an orbital doppler shift ( @xmath90 ) of only 75 km s@xmath54 , well within the range of typical lmxb orbital parameters . however , the highest frequency of the oscillations in burst 2 is greater than the burst 9 at a 3.2@xmath73 level of confidence , and requires a relative doppler shift of 340@xmath96100 km s@xmath54 if the we measure in bursts 1 and 9 in fact reflects the spin frequency of the neutron star . a 300 km s@xmath54 doppler shift is not inconceivable for an lmxb , but ( assuming an inclination of @xmath111 as suggested by the lack of eclipses or dips , and a 1.4 neutron star with a 1 companion ) would suggest a 1 h orbit for the most probable viewing geometry of the bursts placing in the shortest 10% of the lmxb orbital period distribution @xcite . unfortunately no optical companion to has been found to date @xcite , and no orbital period is known for the binary . more information about the system , or a greatly expanded sample of bursts , will be necessary before the binary parameters can be constrained . the fractional change of the oscillation frequency during bursts 1 , 2 , and 9 ( figure [ fev ] ) is within a factor of a few of the changes observed in , and @xcite , @xcite . under the assumptions of the hot spot model , the fractional change in the frequency is directly proportional to the change in the radius at which the hot spot is observed . when angular momentum is conserved , @xmath112 , which for small @xmath91 and @xmath113 gives @xmath114 . for a 1 hz frequency drift corresponds to a @xmath115 of @xmath116 , or about 20 m for a neutron star with a 10 km radius . comparing the various time and length scales which we can infer for the bursts provides some insight into the origin of the oscillations . the change in the radius of the hot spot during the tails of the bursts ( @xmath117 m ) is thought to just arise from the hydrostatic expansion of the nuclear burning layer @xcite . however , the time for the nuclear energy released at the burning layer to diffuse through the neutron star atmosphere to the photosphere during a burst can also be as large as several seconds @xcite . if the hot spot is moving with respect to the surface of the neutron star ( and the atmosphere ) more quickly than the signal from the hot spot can escape the atmosphere , the signal will be unobservable . cumming & bildsten ( 2000 ) have shown that the smaller opacity of a helium rich shell reduces the thermal diffusion time through a helium rich atmosphere . in addition , they found that the time for the hot spot to drift relative to the surface of the neutron star ( estimated from the inverse of the predicted frequency change ) increases when more helium is present , because the higher mean molecular weight of helium prevents the burning layer from expanding as far . the combination of the two effects implies that , independent of the cause of the hotspot , the signal is more likely to be observed during a pure helium burst . in burst 8 we observe an episode of spin - down during the burst peak that is roughly comparable in duration and magnitude to the spin - down episode observed during @xcite burst a from ( also discussed in @xcite ) ; see figure [ fev](d ) . in this case , however , the spin - down is not accompanied by any evidence for late - time radius expansion , as was observed for strohmayer s burst a ; on the contrary , the entire spin - down episode occurs after the photospheric radius has contracted and settled to its minimum value of 8 km . this observation presents a challenge to the spin - frequency interpretation . explaining the spin - down during burst 8 under this model seems to require postulating a relatively consistent m s@xmath54 expansion of the burning layer over the s interval of the pulsations , without any corresponding spectral evidence for this expansion . naturally , a small expansion like this would not be seen directly in the blackbody radius of the spectral fits . however , some additional input of energy , relative to the more typical spin - up seen in bursts 1 , 2 , and 9 , is presumably required to power this expansion , and we see no evidence of this . the amplitude of the burst oscillations we observe increases with photon energy ( figure [ oscvse ] ) in a manner consistent with general - relativistic models of hot spots on the surface of a rotating neutron star @xcite . these models allow larger oscillation amplitudes ( 2080% for two hot spots , as the photon energy increases from 020 kev ) than we observe , but this can be accounted for by allowing the hot spots on to have a finite area on the neutron star surface , which will dampen the modulation ( the intent of miller & lamb 1998 was to calculate upper limits on the pulse amplitudes , so point - source hot spots were assumed ) . phase connection of the pulsations in the bursts from allows us to make sensitive tests for the presence of harmonics and sub - harmonics of the main signal . we find marginal evidence ( 2.1@xmath73 level ) for modulation at the first harmonic of the main 524 hz signal with an average power 1.9@xmath95% that of the main signal , and no evidence ( @xmath995% confidence ) for modulation at frequencies of 1/2 or 3/2 the frequency of the main signal ; our 2@xmath73 upper limits for the power of these components are 5.5% and 4.4% the power of the main signal , respectively . the sub - harmonic modulation in found during oscillations in the burst rise has an average strength 19% that of the main signal @xcite , and is compelling evidence that the spin frequency of that neutron star is in fact half of the strongest burst frequency . this would be expected in beat frequency models , since @xmath118 . the strongest observed oscillations in are in the burst tail , so it is possible that the small upper limit on the sub - harmonic content observed from this source ( by a factor of ) reflects greater symmetry in the hot spots present in the burst tail . however , the evidence we find for modulation at the first harmonic of 524 hz suggests that it is more likely that @xmath119 hz in , since % harmonic content might easily be produced by a pulse which deviates slightly from a sinusoidal profile . should observations of further bursts from strengthen the evidence for a harmonic , the case for @xmath120 in this source will be strengthened , and the beat frequency interpretation cast into doubt . we have observed nine type i x - ray bursts from with the _ rxte_pca . we find that the bursts can be separated into two categories ( section 2.2 , 2.3 ) : fast bursts occur on the banana branch when the accretion rate is high , and generally exhibit short decay time scales , high peak flux , radius expansion episodes , and coherent oscillations . slow bursts occur in the island state at lower accretion rates , and have lower peak fluxes , higher fluences , and longer decay time scales , but exhibit no evidence of either radius expansion or coherent oscillations . the fast and slow bursts may occur when the nuclear fuel is rich in helium and hydrogen respectively ( sections 1.1 and 3.1 ) . we have analyzed the frequency evolution of the pulsations during the bursts by applying the technique of pulsar phase connection , which allows us to `` count pulses '' and thereby characterize the evolution uniquely ( section 2.5 ) . we find that the frequency evolution during the peak and decay of the bursts exhibits a variety of behaviors , including a sharp spin - down in one burst ( burst 8) which challenges the attribution of the frequency evolution during bursts to expansion or contraction of an expanded burning layer . the frequency evolution during bursts 1 and 9 , separated by 2.6 years , extends for over 9 s in both cases and is consistent with nearly the same exponential - relaxation model , with a best - fit frequency difference of [email protected] hz . the evolution of the oscillations in burst 2 is more complicated , and the maximum frequency of the oscillation is greater than those modeled with exponentials by [email protected] hz . phase connection of the burst pulsations allows us to make several precision tests of their properties ( section 2.6 ) . we demonstrate the coherence of these pulsations ( after accounting for their gross frequency evolution ) for @xmath1215000 cycles . we find that the pulsations are spectrally hard in comparison to the burst emission , with the strength of the pulsations increasing monotonically with photon energy . coherently summing the signals from all the bursts with detected pulsations , we find marginal ( 2.1@xmath73 ) evidence for modulation at the first harmonic of the main pulse frequency at 524 hz , with an average power equivalent to 1.9@xmath95% that of the main signal . in combination with our upper limits on modulation at 1/2 or 3/2 the main pulse frequency ( 5.5% and 4.3% the power of the main signal , respectively ) , this suggests that the spin frequency in is more likely to be than half that , as the measured frequency difference of the khz qpos ( @xmath122 hz ) might otherwise suggest . we thank don a. smith for providing us with the data from his _ rxte_proposals for , and deepto chakrabarty for pointing us towards phase connection as the best way to model the frequency evolution during bursts ( and teaching us how to do it ) . this research was partially supported by nasa via grants nag 5 - 8658 and nagw-4517 and by the national science foundation under grant no . phy94 - 07194 . l. b. is a cottrell scholar of the research corporation . ayasli , s. & joss , p. c. 1982 , , 256 , 637 barret , d. 1992 , , 394 , 615 barret , d. , motch , c. , & predehl , p. 1998 , a&a , 329 , 965 barret , d. , olvie , j. f. , boirin , l. , done , c. , skinner , g. k. , & grindlay , j. e. 1999 , astro - ph/9911042 basinska , e. m. , lewin , w. h. g. , sztajno , m. , cominsky , l. r. , & marshall , f. j. 1984 , , 281 , 337 bildsten , l. 1995 , , 438 , 852 bildsten , l. 1997 , _ the many faces of neutron stars _ , r. buccheri , j. van paradijs & m.a . alpar ( eds ) , nato asi series c , vol . 515 , pp . 337 - 368 ( kluwer academic publishers ) , astro - ph/9709094 bildsten , l. 2000 , to appear in `` cosmic explosion '' eds . s. s. holt and w. w. zhang ( aip : new york ) , astro - ph/0001135 bradt , h. v. d. , rothschild , r. e. , & swank , j. h. 1993 , a&as , 97 , 355 chakrabarty , d. 1996 , ph.d . thesis , california institute of technology clark , g. w. , li , f. k. , canizares , c. , hayakawa , s. , jernigan , g. , & lewin , w. h. g. 1977 , , 179 , 651 cocchi , m. , int zand , j. j. m. , bazzano , a. , heise , j. , muller , j. , natalucci , l. , & ubertini , p. 1998 , nuclear physics b ( proc . suppl . ) , 69/1 - 3 , 232 cumming , a. & bildsten , l. 2000 , submitted to deeter , j. e. , pravdo , s. h. , boynton , p. e. 1981 , apj , 247 , 1003 ford , e. c. 1999 , , 519 , l73 fujimoto , m. y. , hanawa , t. , & miyaji , s. 1981 , , 246 , 267 fujimoto , m. y. , sztajno , m. , lewin , w. h. g. , & van paradijs , j. 1987 , , 319 , 902 fushiki , i. & lamb , d. q. 1987 , , 323 , l55 groth , e. j. 1975 , apjs , 29 , 285 gottwald , m. , haberl , f. , parmar , a. n. , & white , n. e. 1986 , , 308 , 213 hanawa , t. , & fujimoto , m. y. 1982 , , 34 , 495 hanawa , t. , & fujimoto , m. y. 1984 , , 36 , 199 jahoda , k. , swank , j. h. , giles , a. b. , stark , m. j. , strohmayer , t. , zhang , w. , & morgan , e. h. 1996 , spie , 2808 , 59 kaspi , v. m. , chakrabarty , d. , & steinberger , j. , 1999 , apj , 525 , 33 langmeier , a. , sztajno , m. , hasinger , g. , trmper , j. , & gottwald , m. 1987 , , 323 , 288 leahy , d. a. , darbro , w. , elsner , r. f. , weisskopf , m. c. , sutherland , p. g. , kahn , s. , & grindlay , j. e. 1983 , , 266 , 160 lewin , w. h. g. , van paradijs , j. , cominsky , l. , & holzner , s. 1980 , , 193 , 15 lewin , w. h. g. , van paradijs , j. , & taam , r. e. 1993 , space science review , 62 , 223 lewin , w. h. g. , pennix , w. van paradijs , j. , damen , e. , sztajno , m. , trmper , j. , & van der klis , m. 1987 , , 319 , 893 london , r. a. , taam , r. e. , & howard , w. m. 1984 , , 287 , l27 london , r. a. , taam , r. e. , & howard , w. m. 1986 , , 306 , 170 kuulkers , e. , & van der klis , m. , 1995 , , 450 , 748 makishima 1983 , 267 , 310 manchester , r. n. , & taylor , j. h. 1977 , pulsars ( san fracisco : w. h. freeman and co. ) markwardt , c. b. , strohmayer , t. e. , & swank , j. h. 1999 , , 512 , l125 mndez , m. & van der klis , m. 1999 , , 517 , l51 mndez , m. , van der klis , m. , & van paradijs , j. 1998 , , 506 , l117 mndez , m. , van der klis , m. , wijndands , r. , ford , e. c. , van paradijs , j. , & vaughan , b. a. 1998 , , 505 , l23 middleditch , j. 1975 , ph.d . thesis , university of california , berkeley , lbl report no . 3639 miller , m. c. 1999 , , 515 , l77 miller , m. c. 2000 , , submitted ( astro - ph 9904093 ) miller , m. c. & lamb , f. k. 1998 , , 499 , l37 miller , m. c. , lamb , f. k. , & psaltis , d. 1998 , , 508 , 791 murakami 1980 , , 240 , l143 osherovich , v. & titarchuk , l. 1999 , 522 , l113 predehl , p. & schmitt , j. h. m. m. 1995 , a&a , 293 , 889 press , w. h. , teukolsky , s. a. , vetterling , w. t. & flannery , b. p. 1992 , numerical recipes in c , 2nd ed . ( cambridge : cambridge university press ) psaltis , d. 1998 , , 501 , l95 smith , d. a. , morgan , e. h. , & bradt , h. 1997 , , 479 , l137 stella , l. , & vietri , m. , 1999 , phys rev lett , 82 , 17 strohmayer , t. e. 1999 , , 523 , l51 strohmayer , t. e. & markwardt , c. b. 1999 , , 516 , l81 strohmayer , t. e. , swank , j. h. , & zhang , w. 1998 , nuclear physics b ( proc . suppl . ) , 69/1 - 3 , 129 strohmayer , t. e. , zhang , w. , & swank , j. h. 1997 , , 487 , l77 strohmayer , t. e. , zhang , w. , swank , j. h. , white , n. e. , & lapidus , i. 1998 , , 498 , l135 strohmayer , t. e. , zhang , w. , swank , j. h. , smale , a. , titarchuk , l. , & day , c. 1996 , 469 , l9 strohmayer , t. e. , zhang , w. , swank , j. h. & lapidus , i. 1998 , , 503 , l147 sunyaev , r. 1989 , iau circ . , 4839 sunyaev , r. 1990 , pisma astron . , 16 , 136 sztajno , m. , bashinka , e. m. , cominsky , l. r. , marshall , f. j. , & lewin , w. h. g. 1983 , , 267 , 713 sztajno , m. , van paradijs , j. , lewin , w. h. g. , trmper , j. , stollman , g. , pietsch , w. , & van der klis , m. 1985 , , 299 , 487 sztajno , m. , van paradijs , j. , lewin , w. h. g. , langmeier , a. , trmper , j. , & pietsch , w. 1986 , , 222 , 499 van der klis , m. , hasinger , g. , damen , e. , pennix , w. , paradijs , j. , & lewin , w. h. g. 1990 , , 360 , l19 van der klis , m. 1995 , in _ x - ray binaries _ , p. 252 , lewin , w. h. g. , van paradijs , j. , & van den heuvel , e. p. j. ( eds ) , cambridge university press van der klis , m. 1997 in _ the many faces of neutron stars _ , r. buccheri , j. van paradijs & m.a . alpar ( eds ) , nato asi series c , vol . 515 , pp . 337 - 368 ( kluwer academic publishers ) , astro - ph/9710016 in the proceedings of the third william fairbank meeting , rome , astro - ph/9812395 van der klis , m. 2000 , to appear in the annual review of astronomy and astrophysics , astro - ph/0001167 van der klis , m. , wijnands , r. a. d. , horne , k. , chen , w. 1997 , 481 , l97 van paradijs , j. , pennix , w. , & lewin , w. h. g. 1988 , , 233 , 437 van paradijs , j. , pennix , w. , lewin , w. h. g. , sztajno , m. , & trmper , j. 1988 , a&a , 192 , 147 vaughan , b. a. et al . 1994 , , 435 , 362 white , n. e. , nagase , f. , & parmar , a. n. 1995 , _ x - ray binaries _ , p. 1 , lewin , w. h. g. , van paradijs , j. , & van den heuvel , e. p. j. ( eds ) , cambridge university press wijnands , r. a. d. & van der klis , m. 1997 , , 482 , l65 zhang , w. , jahoda , k. , kelley , r. l. , strohmayer , t. e. , swank , j. h. , & zhang , s. n. 1998 , , 495 , l9 lcccccccc 10416 - 01 - 01 - 00 & 1996 jul 14 & 02:32:31 & 07:05:31 & 9960 & 552 & 6.38 & 1.63 & 0.40 + 10416 - 01 - 02 - 00 & 1996 aug 01 & 16:53:56 & 20:51:56 & 9900 & 344 & 2.74 & 1.50 & 0.39 + 10416 - 01 - 03 - 00 & 1996 aug 31 & 17:40:17 & 19:44:16 & 3540 & 405 & 4.73 & 1.54 & 0.38 + 20085 - 01 - 01 - 01 & 1997 oct 28 & 22:12:38 & 03:41:39 & 13080 & 404 & 4.69 & 1.62 & 0.40 + 20085 - 01 - 01 - 02 & 1997 oct 29 & 06:22:45 & 08:28:45 & 4320 & 467 & 5.37 & 1.65 & 0.41 + 20085 - 01 - 01 - 03 & 1997 oct 29 & 18:00:34 & 20:00:33 & 4320 & 489 & 5.63 & 1.62 & 0.43 + 20085 - 01 - 01 - 00 & 1997 oct 29 & 21:07:33 & 11:39:33 & 28260 & 505 & 5.85 & 1.71 & 0.46 + 30061 - 01 - 01 - 00 & 1998 jul 31 & 11:26:08 & 15:17:07 & 9000 & 377 & 4.36 & 1.65 & 0.40 + 30061 - 01 - 01 - 01 & 1998 aug 01 & 01:49:06 & 04:09:07 & 6480 & 381 & 4.42 & 1.53 & 0.42 + 30061 - 01 - 01 - 02 & 1998 aug 01 & 06:47:02 & 17:18:02 & 22980 & 369 & 4.28 & 1.63 & 0.41 + 30061 - 01 - 01 - 03 & 1998 aug 01 & 23:04:02 & 02:54:02 & 9480 & 369 & 4.28 & 1.64 & 0.41 + 30061 - 01 - 01 - 04 & 1998 aug 02 & 08:32:00 & 09:07:56 & 2160 & 378 & 4.36 & 1.65 & 0.40 + 30061 - 01 - 02 - 00 & 1998 oct 02 & 03:26:02 & 08:40:02 & 11040 & 106 & 1.26 & 1.73 & 0.72 + 30061 - 01 - 02 - 01 & 1998 oct 02 & 13:01:58 & 14:01:58 & 3600 & 118 & 1.41 & 1.74 & 0.73 + 30061 - 01 - 02 - 02 & 1998 oct 03 & 01:49:53 & 04:11:54 & 5100 & 103 & 1.23 & 1.73 & 0.71 + 30061 - 01 - 02 - 03 & 1998 oct 03 & 05:44:53 & 09:14:53 & 7800 & 109 & 1.30 & 1.61 & 0.71 + 30061 - 01 - 02 - 04 & 1998 oct 04 & 05:35:44 & 10:49:44 & 11880 & 107 & 1.29 & 1.69 & 0.71 + 30061 - 01 - 02 - 05 & 1998 oct 05 & 16:13:32 & 17:12:33 & 3540 & 108 & 1.30 & 1.74 & 0.73 + 30061 - 01 - 02 - 06 & 1998 oct 06 & 01:49:29 & 09:11:29 & 15360 & 115 & 1.37 & 1.73 & 0.72 + 30061 - 01 - 02 - 07 & 1998 oct 08 & 10:12:10 & 10:47:10 & 2100 & 128 & 1.55 & 1.58 & 0.61 + 30061 - 01 - 03 - 00 & 1999 feb 22 & 18:23:17 & 20:59:17 & 7380 & 271 & 3.22 & 1.46 & 0.42 + 40409 - 01 - 01 - 00 & 1999 feb 22 & 23:46:17 & 05:00:17 & 9300 & 266 & 3.16 & 1.56 & 0.42 + 30061 - 01 - 03 - 01 & 1999 feb 25 & 21:35:42 & 23:57:42 & 6060 & 322 & 3.76 & 1.62 & 0.41 + 30061 - 01 - 03 - 02 & 1999 feb 25 & 23:59:42 & 00:12:42 & 780 & 313 & 3.65 & 1.62 & 0.41 + 30061 - 01 - 04 - 00 & 1999 feb 26 & 15:08:47 & 20:58:48 & 13380 & 390 & 4.55 & 1.69 & 0.45 + 30061 - 01 - 04 - 01 & 1999 feb 27 & 11:56:54 & 14:23:55 & 6840 & 389 & 4.51 & 1.61 & 0.44 + 30061 - 01 - 04 - 02 & 1999 feb 27 & 15:08:56 & 17:35:56 & 6120 & 380 & 4.38 & 1.69 & 0.42 + clccccccc 1 & 1996 jul 14 04:23:42 & 3.25 & 4.0 & 4.08(8 ) & 4.6(1 ) & & & + & & & 11.0 & 9.4(6 ) & & 3.24(7 ) & 6.6(2 ) & y + 2 & 1997 oct 29 08:10:23 & 3.00 & 4.5 & 4.4(1 ) & 6.3(1 ) & & & + & & & 10.0 & 11.4(3 ) & & 4.25(7 ) & 6.8(2 ) & y + 3 & 1998 oct 02 07:34:54 & 3.75 & 9.0 & 11.0(1 ) & 4.4(1 ) & 7.01(7 ) & 15.9(4 ) & n + 4 & 1998 oct 02 14:00:27 & 3.50 & 9.0 & 7.6(2 ) & 4.7(1 ) & & & + & & & 20.0 & 9.4(2 ) & & 7.02(8 ) & 14.6(4 ) & n + 5 & 1998 oct 03 09:11:55 & 5.25 & 9.0 & 11.5(1 ) & 3.4(1 ) & 5.85(7 ) & 17.2(5 ) & n + 6 & 1998 oct 06 03:26:58 & 2.75 & 9.0 & 9.11(6 ) & 4.4(1 ) & 6.78(7 ) & 15.1(4 ) & n + 7 & 1999 feb 23 03:06:16 & 1.25 & 1.5 & 3.7(1 ) & 2.1(1 ) & 0.77(2 ) & 3.3(2 ) & n + 8 & 1999 feb 26 17:10:53 & 1.50 & 5.0 & 4.01(9 ) & 5.3(1 ) & & & + & & & 10.0 & 9.2(2 ) & & 4.57(7 ) & 8.6(2 ) & y + 9 & 1999 feb 27 17:23:01 & 2.00 & 2.5 & 6.27(7 ) & 5.2(1 ) & & & + & & & 15.0 & 14.6(9 ) & & 4.65(9 ) & 8.6(2 ) & y + lcccccc 1 & @xmath123 & @xmath8311.1 & @xmath835.9 & 2 & 3 & y + 2 & @xmath124 & @xmath836.9 & @xmath836.1 & 4 & 2 & y + 3 & @xmath839.2 & @xmath839.2 & @xmath839.9 & & & n + 4 & @xmath836.6 & @xmath838.0 & @xmath838.7 & & & n + 5 & @xmath838.9 & @xmath837.8 & @xmath839.0 & & & n + 6 & @xmath838.0 & @xmath837.2 & @xmath836.7 & & & n + 7 & @xmath125 & @xmath8313.3 & @xmath838.9 & 0 & 1 & n + 8 & @xmath126 & @xmath835.6 & @xmath8311.5 & 3 & 1 & y + 9 & @xmath127 & @xmath835.6 & @xmath837.8 & 2 & 6 & y + ccrrrrrc 1 & 2.0011.75 & [email protected] & [email protected] & [email protected] & @[email protected] & & 9.5/8 + 2 & 2.510.5 & [email protected] & [email protected] & @[email protected] & [email protected] & @[email protected] & 5.0/5 + 7 & 0.251.00 & [email protected] & [email protected] & @xmath834.5 & & & 7e-5/1 + 8 & 0.000.75 & [email protected] & [email protected] & @xmath835.4 & & & 0.01/1 + 8 & 2.05.0 & [email protected] & [email protected] & @[email protected] & @xmath834.6 & & 7.0/2 + 9 & 2.011.0 & [email protected] & [email protected] & [email protected] & @[email protected] & & 13.3/11 +	we present an analysis of the full set of _ rossi x - ray timing explorer _ ( _ rxte _ ) observations of made between 1996 august and 1999 february , concentrating on the nine type i x - ray bursts that were observed . we find that the bursts divide naturally into two populations : `` fast bursts '' occur on the banana branch when the accretion rate is high and exhibit short decay times , high peak fluxes , and radius expansion episodes . `` slow bursts '' occur in the island state at lower accretion rates , have lower peak fluxes , higher fluences , longer decay times , and show no evidence of radius expansion . all five of the fast bursts , and none of the four slow bursts , show coherent oscillations near 524 hz . thus the mechanism that produces the burst pulsations may well be related to the helium - rich burning process indicated by the other properties of the fast bursts . we perform in - burst phase connection of the burst pulsations , which allows us to unambiguously characterize their frequency evolution . that evolution exhibits a variety of behaviors , including a sharp spin - down during one burst . the frequency evolution exhibited by two bursts that occurred 2.6 years apart is consistent with nearly the same exponential - relaxation model ; in particular , the asymptotic frequencies of the two bursts differ by only @xmath0 hz . the evolution during another burst , not modeled as exponential , shows a maximum frequency which is @xmath1 hz greater than the smaller of these asymptotic frequencies . applying our phase models , we find that the pulsations are spectrally harder than the burst emission , with the strength of the pulsations increasing monotonically with photon energy . coherently summing all of the burst signals , we find upper limits of 5% relative to the power of the main pulse on any modulation at 1/2 or 3/2 the main pulse frequency . we discuss our results in the context of models in which the burst pulsations are produced by a hot spot on the surface of a spinning neutron star .
nowadays there is wide consensus that most if not all nearby early type galaxies host a supermassive black hole ( smbh ) at their centers ( see for a recent review ferrarese 2006 ) . however a direct determination of the bh mass ( @xmath6 ) is possible only for a limited number of inactive galaxies in the local universe , from dynamical methods ( virial theorem ) based on the observations of the kinematics of stars orbiting around the smbh . an extension of this method to higher redshift objects is viable only for active galaxies , where @xmath6 can be derived from the dynamics of gas gravitationally bound to the smbh : this is the usual case of regions emitting the broad emission lines ( blr ) . at low redshift correlations between @xmath6 and the global properties of the host galaxies , namely the luminosity ( @xmath7 ) and the central velocity dispersion of the bulge , have been found . locally therefore the measure of the properties of the host galaxy is an alternative method for evaluating @xmath6 . the aim of this paper is to compare the mass determinations through the virial method and through the global parameters of the hosts of close by quasars . for qsos it is very difficult to measure the velocity dispersion of the stars in the bulge , so one must rely on the @xmath8 relationship . for nearby ellipticals , there is increasing evidence that this relation has the same behavior for both inactive and active galaxies ( see for example merritt & ferrarese 2001 ; mclure & dunlop 2001 , 2002 ; bettoni et al . 2003 ) , notwithstanding a substantial increase of the dispersion in the latter case , possibly due to the greater difficulty in the determination of all the parameters . moreover , agn hosts and inactive galaxies appear substantially the same objects , as indicated by direct morphological studies ( e.g. dunlop et al . this is consistent with a picture where the agn phenomenon , which is linked to the gas accretion on the central bh , would be a phase characterizing the life of normal galaxies in particular during major mergers ( see for example kauffmann & haehnelt 2000 ) . for these reasons , in this work we start with the assumption that the correlation established for nearby galaxies is also valid for higher redshift qsos , provided that host galaxy luminosity is corrected for galactic evolution to @xmath9 . a second set of qso bh mass determinations can be obtained through the virial theorem , provided that an estimate of the velocity @xmath10 and radius @xmath11 of the blr is available . the blr velocity can be derived from the fwhm of the broad emission lines if one assumes a value for the geometrical factor @xmath12 , which is an unknown parameter of order unity linked to the geometry of the blr . the measurement of @xmath11 may require cumbersome procedures , like reverberation mapping ( ulrich et al . 1984 ; blandford & mckee 1982 ; peterson 2001 ) , that uses the continuum source variability to evaluate time delays between the central emission and the blr gas response ; an alternative is to infer the blr size from the ionizing luminosity with which it is correlated ( kaspi et al . 2000 ; vestergaard 2002 ; pian , falomo & treves 2005 ; kaspi et al . 2005 ; vestergaard & peterson 2006 ) . bh mass determinations can be then derived from a measure of the broad line widths and of the ionizing continuum intensity , once the value of the geometrical factor is fixed . in the next section we focus on the quasar sample choice and on the data collection and analysis ( [ secima][secspe ] ) . in section [ secmas ] we describe the formulae used for applying the two methods of mass determination ( [ secpho][secvir ] ) and report on the results ( [ seccomp ] ) . this analysis constrains the value of the geometrical factor . the issue of the blr geometry and orientation is further discussed in paragraph [ calcolof ] . in section [ secsum ] our results are compared with those of literature and discussed within the unified model of agn . throughout this paper , we adopt a concordant cosmology with @xmath13 km s@xmath14 mpc@xmath14 , @xmath15 and @xmath16 . we have converted the results of other authors to this cosmology when adopting their relations and using their data . we selected all qsos from the veron cetty & veron ( 2003 ) catalogue , for which _ hst _ images and spectra are available . we further require that the images have exposure times longer than 1000 s and that spectra cover the range 14001700 ( @xmath17 band ) . in addition we selected only objects hosted by ellipticals , so that the bulge component substantially coincides with the whole galaxy . the resulting sample includes 29 objects at @xmath180.6 , 17 of which are radio loud ( rlq ) and 12 are radio quiet ( rqq ) . the main properties of the objects are listed in table 1 . the sample is essentially uniformly distributed in redshift ( see fig . [ zdistrib ] ) . in table 5 we report the journal of observations for the two objects without available measurement of the host galaxy from the literature . a two dimensional analysis has been carried out using aida ( astronomical image decomposition and analysis , uslenghi & falomo 2006 , in preparation ; see also kotilainen et al . 2005 ) , a software package specifically designed to perform two dimensional model fitting of qso images , providing simultaneous decomposition into the nuclear and host galaxy components . psf modelling is by far the most critical step of analysing qso host galaxies . for hst images , the software package tinytim ( krist & hook 1997 ) can produce a psf model that is very accurate in the inner part ( up to 12 ) but it does not properly model the external fainter halo produced by the scattered light ( e.g. scarpa et al . 2000 ) . to account for this we therefore modelled the external part by adding an exponential term to the tinytim psf . the parameters of these components have been evaluated by fitting the psf model to a number of stars in the field . for each qso we have then applied a mask to exclude possible contamination from spurious sources , bad pixels , companions and other defects affecting the image . the background was estimated from the signal computed in a circular annulus centered on the object and its uncertainty obtained from the comparison of the values computed in annuli at different radii . the masked qso images were finally fitted with a model including both a point source and the host galaxy assuming a de vaucouleurs law convolved with the psf . the results of the best fit parameters are reported in table 5 . a fit with a disc galaxy yielded significant poorer results , consistently with the high luminosity of the nucleus ( see dunlop et al . 2003 ) . in fig . [ rp ] we show the qsos images and residuals . the radial profiles of the images , of the psf , of the galaxy and of the best fitting models for the 2 objects are also shown .	we report on the analysis of the photometric and spectroscopic properties of a sample of 29 low redshift ( @xmath0 ) qsos for which both _ hst _ wfpc2 images and ultraviolet _ hst _ fos spectra are available . for each object we measure the @xmath1 band absolute magnitude of the host galaxy , the civ(1550 ) line width and the 1350 continuum luminosity . from these quantities we can estimate the black hole ( bh ) mass through the @xmath2 relation for inactive galaxies , and from the virial method based on the kinematics of the regions emitting the broad lines . the comparison of the masses derived from the two methods yields information on the geometry of the gas emitting regions bound to the massive bh . the cumulative distribution of the line widths is consistent with that produced by matter laying in planes with inclinations uniformly distributed between @xmath310@xmath4 and @xmath350@xmath4 , which corresponds to a geometrical factor @xmath5 . our results are compared with those of the literature and discussed within the unified model of agn . [ firstpage ] galaxies : active galaxies : nuclei quasars : general .
the omnipresence of power - laws in natural , socio - economic , technical , and living systems has triggered immense research activity to understand their origins . it has become clear in the past decades that there exist several distinct ways to generate power - laws ( or asymptotic power - laws ) , for an overview see for example @xcite . in short , power - laws of the form @xmath0 arise in critical phenomena @xcite , in systems displaying self - organized criticality @xcite , preferential attachment type of processes @xcite , multiplicative processes with constraints @xcite , systems described by generalized entropies @xcite , or sample space reducing processes @xcite , i.e. processes that reduce the number of possible outcomes ( sample space ) as they unfold . literally thousands of physical , natural , man - made , social , and cultural processes exhibit power - laws , the most famous being earthquake magnitudes @xcite , city sizes @xcite , foraging and distribution pattern of various animal species @xcite , evolutionary extinction events @xcite , or the frequency of word occurrences in languages , known as zipf s law @xcite . it is obvious that estimating power - law exponents from data is a task that sometimes should be done with high precision . for example if one wants to determine the universality class a given process belongs to , or when one estimates probabilities of extreme events . in such situations small errors in the estimation of exponents may lead to dramatically wrong predictions with potentially serious consequences . estimating power - law exponents from data is not an entirely trivial task . many reported power - laws are simply not exact power - laws , but follow other distribution functions . despite the importance of developing adequate methods for distinguishing real power - laws from alternative hypotheses , we will not address this issue here since good standard literature on the topic of bayesian _ alternative hypotheses testing _ exists , see for example @xcite . for power - laws some of these matters have been discussed also in @xcite . here we simply focus on estimating power - law exponents from data on a sound probabilistic basis , using a classic bayesian parameter estimation approach , see e.g. @xcite , that provides us with _ maximum likelihood _ ( ml ) estimators for estimating power - law exponents over the full range of reasonably accessible values . having such estimators is of particular interest for a large classes of situations where exponents close to @xmath1 appear ( zipf s law ) . we will argue here that whenever dealing with data we can assume discrete and bounded samples spaces ( domains ) , which guarantees that power - laws are normalizable for arbitrary powers @xmath2 . we then show that the corresponding ml estimator can then also be used to estimate exponents from data that is sampled from continuous sample spaces , or from sample spaces that are not bounded from above . in physics the theoretical understanding of a process sometimes provides us with the luxury of knowing the exact form of the distribution function that one has to fit to the data . for instance think of critical phenomena such as ising magnets in 2 dimensions at the critical temperature , where it is understood that the susceptibility follows a power - law of the form @xmath3 , with @xmath4 a critical exponent , that occasionally even can be predicted mathematically . however , often and especially when dealing with complex systems we do not enjoy this luxury and usually do not know the exact functions to fit to the data . in such a case , let us imagine that you have a data set and from first inspection you think that a power - law fit could be a reasonable thing to do . it is then essential , before starting with the fitting procedures , to clarify what one knows about the process that generated this data . the following questions may help to do so . plfit and r@xmath5plhistfit can be used . [ fig : diagram ] ] * do you have information about the dynamics of the process that is generating what appears to be a power - law ? * is the data generated by a bernoulli process ( e.g. tossing dice ) , or not ( e.g. preferential attachment ) ? * is the data available as a collection of samples ( a list of measurements ) , or only coarse - grained in form of a histogram ( binned or aggregated data ) . * is the data sampled from a discrete ( e.g. text ) or continuous sample space ( e.g. earthquakes ) ? * does the data have a natural ordering ( e.g. magnitudes of earthquakes ) , or not ( e.g. word frequencies in texts ) ? the decisions one has to take before starting to estimate power - law exponents are shown as a decision - tree in fig . ( [ fig : diagram ] ) . if it is known that the process generating the data is not a bernoulli process ( for example if the process belongs to the family of history dependent processes such as e.g. preferential attachment ) , then one has the chance to use this information for deriving parameter estimators that are tailored exactly for the particular family of processes . if no such detailed information is available one can only treat the process as if it were a bernoulli process , i.e. information about correlations between samples is ignored . if we know ( or assume ) that the data generation process is a bernoulli process , the next thing to determine is whether the data is available as a collection of data points , or merely as coarse grained information in form of a histogram that collects distinct events into bins ( e.g. histograms of logarithmically binned data ) . if data is available in form of a data set of samples ( not binned ) , a surprisingly general maximum likelihood ( ml ) estimator can be used to predict the exponent of an underlying power - law @xmath6 . this estimator that we refer to as @xmath7 , will be derived in the main section . its estimates for the underlying exponent @xmath2 , are denoted by @xmath8 . the code for the corresponding algorithm we refer to as ` r_plfit ` . if information is available in form of a histogram of binned data , a different estimator becomes necessary . the corresponding algorithm ( ` r_plhistfit ` ) is discussed in appendix a and in the section below on discrete and continuous sample spaces . both algorithms are available as matlab code @xcite . for how to use these algorithms , see appendix b. if we have a dataset of samples ( not binned ) , so that the ` r_plfit ` algorithm can be used , it still has to be clarified whether the data has a natural order or not ? numerical observables such as earthquake magnitudes are _ naturally _ ordered . one earthquake is always stronger or smaller than the other . if observables are non - numeric , such as word types in a text , then a natural order can not be known _ a priori_. the natural order can only be inferred approximately by using so - called _ rank - ordering _ ; or alternatively by using the so - called _ frequency distribution _ of the data . details are discussed below in the section on rank - order , frequency distributions , and natural order . other issues to clarify are to see if a given sample space is continuous or discrete , and if the sample space is bounded or unbounded . these questions however , turn out to be not critical . one might immediately argue that for unbounded power - law distribution functions normalization becomes an issue for exponents @xmath9 . however , this is only true for bernoulli processes on _ unbounded _ sample spaces . since all real - world data sets are collections of finite discrete values one never has to actually deal with normalization problems . moreover , since most experiments are performed with apparati with finite resolution , most data can be treated as being sampled from a bounded , discrete sample space , or as binned data . for truly continuous processes the probability of two sampled values being identical is zero . therefore , data sampled from continuous distributions can be recognized by sample values that are unique in a data set . see appendix a for more details . statistically sound ways to fit power - laws were advocated and discussed in @xcite . they overcome intrinsic limitations of the _ least square _ ( ls ) fits to logarithmically scaled data , which were and are widely ( and often naively ) used for estimating exponents . the ml estimator that was presented in @xcite we refer to as the @xmath10 ( for clauset - shalizi - newman ) estimator ; its estimates for the exponent we denote by @xmath11 . the approach that leads to @xmath10 focuses on continuous data @xmath12 that follows a power - law distribution from eq . ( [ pow ] ) , and that is bounded from below @xmath13 but is not bounded from above ( i.e. @xmath14 with @xmath15 ) . in @xcite emphasis is put on how ml estimators can be used to infer whether an observed distribution function is likely to be a power - law or not . also the pros and cons of using cumulative distribution functions for ml estimates are discussed , together with ways of treating discrete data as continuous data . for the continuous and unbounded case , simple explicit equations for the @xmath10 estimator can be derived . the continuous approach however , even though it seemingly simplifies computations , introduces unnecessary self - imposed limitations with respect to the range of exponents that can be reliably estimated . @xmath10 works brilliantly for a range of exponents between @xmath16 and @xmath17 . here we show how to overcome these limitations and by doing so extend the accessible range of exponents by presenting the exact methodology for estimating @xmath2 for discrete bounded data with the estimator @xmath7 . while this approach appears to be more constrained than the continuous one we can show also theoretically that data from continuous and potentially unbounded sample spaces can be handled within essentially the same general ml framework as well . the key to the @xmath7 estimator is that it is not necessary to derive explicit equations for finding @xmath8 . implicit equations in @xmath2 exist for power - law probability distributions over discrete or continuous sample spaces that are both bounded from below _ and _ above . solutions @xmath8 can be easily obtained numerically . an implementation of the respective algorithms can be found in @xcite , for a tutorial see appendix b. there exist three distinct types of distribution functions that are of interest in the context of estimating power - law exponents : + i : : the _ probability distribution _ @xmath18 assigns a probability to every observable state - value @xmath12 . discrete and bounded sample spaces are characterized by @xmath19 state - types @xmath20 , with each type @xmath21 being associated with a distinct value @xmath22 . ii : : the _ relative frequencies _ , @xmath23 , where @xmath24 is the number of times that state - type @xmath21 is observed in @xmath25 experiments . @xmath26 is the _ histogram _ of the data . as explained below in detail , the relative frequencies can be ordered in two ways . + @xmath27 if @xmath28 is ordered according to their descending magnitude this is called the _ rank ordered _ distribution . + @xmath27 if @xmath28 is ordered according to the descending magnitude of the probability distribution @xmath29 , then they are _ naturally ordered _ relative frequencies . iii : : the _ frequency distribution _ @xmath30 counts how many state - types @xmath21 fulfill the condition @xmath31 . in fig . ( [ fig : rankvsfreq ] ) we show these distribution functions . there @xmath32 data points are sampled from @xmath33 , with probabilities @xmath34 . the probability distribution is shown ( red ) . the relative frequency distribution @xmath35 is plotted in natural order ( blue ) , the rank - ordered distribution is shown with the yellow line , which clearly exhibits an exponential decay towards the the tail . the inset shows the frequency distribution @xmath30 of the same data . we next discuss how different sampling processes can be characterized in terms of natural order , rank - order , or frequency distributions . for some sampling processes the ordering of the observed states for example think of @xmath12 representing the numerical values of earthquake magnitudes . here any two observations @xmath12 and @xmath36 can be ordered with respect to their numerical value , or their _ natural order_. since power - law distributions @xmath6 are monotonic this is equivalent to ranking observations according to the probability distribution @xmath37 they are sampled from : the most likely event has _ natural rank _ @xmath38 , the second most likely rank @xmath39 , etc . in other words , we can order state - types @xmath12 in a way that over the sample space @xmath40 , @xmath41 is a monotonic and decreasing function . with an exponent @xmath42 ( red line ) . the relative frequencies @xmath28 are shown for @xmath32 sampled data points according to their natural ( prior ) ordering that is associated with @xmath37 ( blue ) . the rank - ordered distribution ( posterior ) is shown in yellow , where states @xmath21 are ordered according to their observed relative frequencies @xmath28 . the rank - ordered distribution follows a power - law , except for the exponential decay that starts at rank@xmath43 . a low frequency cut - off should be used to remove this part for estimating exponents . the inset shows the frequency distribution @xmath30 that describes how many states @xmath12 appear @xmath44 times ( green ) . the frequency distribution has a maximum and a power - law tail with exponent @xmath45 . to estimate @xmath46 , one should only consider the tail of the frequency distribution function . [ fig : rankvsfreq ] ] if @xmath37 is not known _ a priori _ because the state - types @xmath21 have no numerical values @xmath47 attached , as happens for example with words in a text , we can only count relative frequencies @xmath28 ( a normalized histogram ) of states of type @xmath21 , _ a posteriori _ , i.e. after sampling . to be clear , let @xmath26 be the histogram of @xmath25 recorded states . @xmath24 is the number of times we observed type @xmath21 , then @xmath23 is the relative frequency of observing states of type @xmath21 . after all samples are taken , one can now order states with respect to @xmath28 , such that the rank @xmath38 is assigned to state @xmath21 with the largest @xmath28 , rank @xmath39 to @xmath48 with the second largest @xmath49 , etc . @xmath50 is called the _ rank - ordered _ distribution of the data . the natural order imposed by @xmath37 and the rank - order imposed by @xmath35 are not identical for finite @xmath25 . however , if data points have been sampled independently , then @xmath35 converges toward @xmath37 ( for @xmath51 ) and the rank - order induced by @xmath35 will asymptotically approach the natural order induced by @xmath37 . the highest uncertainty on estimating the order induced by @xmath37 using @xmath35 is associated with the least frequent observations . therefore , when estimating exponents from rank - ordered distributions , one might consider to use a low - frequency cut - off to exclude infrequent data . exponents of power - laws can also be estimated from _ frequency _ distributions @xmath30 . these counts how many distinct state - types @xmath21 occur exactly @xmath44 times in the data . it does not depend on the natural ( prior ) order of states and therefore is sometimes preferred to the ( posterior ) rank - ordered distribution . however , complications may appear also when using @xmath30 . the frequency distribution @xmath30 that is associated with a power - like probability distribution @xmath52 ( and asymptotically to @xmath35 ) is not an exact power - law but a non - monotonic distribution ( with a maximum ) . only its tail decays as a power - law , @xmath53 . the exponents @xmath2 and @xmath46 are related through the well known equation @xmath54 if the probability distribution has exponent @xmath2 , the tail of the associated frequency distribution has exponent @xmath46 . since the frequency distribution behaves like a power - law only in its tail , estimating @xmath46 makes it necessary to constrain the observed data to large values of @xmath44 . note that this is equivalent to using a low - frequency cut - off . one option to do that is to derive a maximum entropy functional for @xmath30 and fit the resulting ( approximate ) max - ent solution to the data . we do not follow this route here . if the natural order of the data is known , one can directly use the natural ordered data in the ml estimates for the exponents . if it is not known , either the rank - ordered distribution can be used to estimate @xmath2 , or the frequency distribution to estimate @xmath46 , see fig . ( [ fig : diagram ] ) . one might also estimate both , @xmath2 in the rank ordered distribution , and @xmath46 in the frequency distribution of the data . ( [ alphalambda ] ) to compare the two estimates may be used as a rough quality - check . if estimates do not reasonably coincide one should check whether the used data ranges have been appropriately chosen . if large discrepancies remain between @xmath46 and @xmath55 this might indicate that the observed distribution function in question is only an approximate power - law , for which eq . ( [ alphalambda ] ) need not hold . for a tutorial on how to use ` r_plfit ` to perform estimates see appendix b. data can originate from continuous sample spaces @xmath56 $ ] , or discrete ones @xmath57 . to each state - type @xmath20 , there is assigned a state - value @xmath47 . whether a distribution function @xmath58 , with @xmath59 , is normalizable or not , can only be decided once the sample space @xmath60 has been specified . the normalization factors for continuous and discrete @xmath60 are @xmath61 for bounded sample spaces with @xmath62 , power - laws are always normalizable for arbitrary exponents @xmath2 , and a well defined ml estimator of @xmath8 exists ( see below ) . the normalization constants in eq . ( [ norm ] ) can be specified in ` r_plfit ` ( see appendix b ) . data sampled from a continuous sample space @xmath63 can essentially be treated as if it were sampled from a discrete sample space @xmath64 , where @xmath65 are given by the unique collection of distinct values in the data set . that is , the data set @xmath66 contains @xmath25 data points @xmath67 ( that have @xmath19 unique values @xmath47 , the states of type @xmath21 ) which we collect in the discrete sample space @xmath68 . for truly continuous data we have @xmath69 , since the probability of @xmath70 for @xmath71 is vanishing . as a consequence the histogram @xmath24 , which counts the number of times @xmath47 appears in the data , is essentially given by @xmath72 for all @xmath20 this provides us with a practical criterion for when to use the normalization constant for discrete or continuous data . for details see appendix a. the equation for the ml estimator @xmath7 , that yields the estimate @xmath8 , only requires the knowledge of the relative frequency distribution @xmath23 ( in natural- or rank - order ) of the observed state - types @xmath21 , as we will see in eq . ( [ bayes4 ] ) below . therefore ` r_plfit ` can work either with data sets @xmath12 or histograms @xmath73 over the unique values in the data sets . if data comes in coarse grained form , i.e. histograms , where each bin may contain a whole range of observable values @xmath12 , then an estimator is required that is different from @xmath7 @xcite , see also appendix a. the corresponding code ` r_plhistfit ` can also be downloaded from @xcite . consider a family of random processes @xmath75 that is characterized by the parameters @xmath76 . let @xmath75 be defined on a discrete sample space @xmath77 , with @xmath78 . the process @xmath75 samples values @xmath79 with probability , @xmath80 let us repeat the process @xmath75 in @xmath25 independent experiments to obtain a data set @xmath81 . @xmath26 is the histogram of the events recorded in @xmath82 , i.e. @xmath24 is the number of times @xmath47 appears in @xmath82 . note that @xmath83 . as a consequence of independent sampling , the probability to sample exactly @xmath73 is , @xmath84 where @xmath85 is the multinomial factor . bayes formula allows us to get an estimator for the parameters @xmath86 , @xmath87 obviously , @xmath88 does not depend on @xmath86 . without further available information we must assume that the parameters @xmath86 are uniformly distributed between their upper and lower limits . as a consequence , @xmath89 also does not depend on @xmath86 within the limits of the parameter range and can be treated as a constant such as @xmath2 does not work for parameters such as @xmath90 and @xmath91 . for those variables it turns out that @xmath89 can not be assumed to be constant between upper and lower bounds of the respective parameter values . bayesian estimators for @xmath90 and @xmath91 require to explicitly consider a non - trivial function @xmath89 . though in principle feasible , we ignore the possibility of deriving bayesian estimates for @xmath90 and @xmath91 in this paper . ] . from eq . ( [ bayes ] ) it follows that the value @xmath92 that maximizes @xmath93 also maximizes @xmath94 . the most likely parameter values @xmath95 are now found by maximizing the log - likelihood , @xmath96 for all parameters @xmath97 . here @xmath98 , is the so - called _ cross - entropy_. in other words , ml - estimates maximize the cross - entropy with respect to the parameters @xmath99 . to apply eq . ( [ bayes3 ] ) for ml - estimates of power - law exponents , one specifies the finite sample space @xmath77 , and the family of probability density functions is , @xmath100 with @xmath59 . note that the set of parameters @xmath86 defined above now only contains @xmath2 , or @xmath101 . the normalization constant is @xmath102 . the derivative with respect to @xmath2 of the cross - entropy , @xmath103 , has to be computed , and setting @xmath104 yields @xmath105 the solution to this implicit equation , @xmath106 , can not be written in closed form but can be easily solved numerically . see @xcite for the corresponding algorithm and appendix b for a tutorial . one possibility to find the solution @xmath106 from the implicit equation eq . ( [ bayes4 ] ) , is to iteratively refine approximate solutions . for this , select @xmath107 values @xmath2 from the interval @xmath108 $ ] , where @xmath109 is a finite fixed number , say @xmath110 . those values may be chosen to be given by the expression @xmath111 for @xmath112 . the parameters @xmath113 and @xmath114 are defined in the following way : first define @xmath115 , and @xmath116 , where @xmath117 and @xmath118 are parameters of the algorithm . then define @xmath119 with @xmath120 . if @xmath121 is the optimal solution of eq . ( [ bayes4 ] ) for some @xmath122 , then we can choose @xmath123 , and @xmath124 and @xmath125 . one then continues by iterating @xmath126 times until @xmath127 , where @xmath128 is the desired accuracy of the estimate of @xmath8 . as a consequence , the value @xmath129 , for which @xmath130 holds , optimally estimates @xmath8 in the @xmath126th iteration with an error smaller than @xmath128 . note that @xmath128 is the error of the @xmath7-estimator with respect to the exact value of the predictor @xmath8 , and is not the error of @xmath8 with respect to the ( typically unknown ) value of the exponent @xmath2 of the sampling distribution . , and @xmath7 . for @xmath131 values of @xmath2 in the range between 0 and 4 , we sample @xmath132 events from @xmath133 , from a power - law probability distribution @xmath134 . the estimated exponents @xmath135 for the estimators @xmath136 ( red ) , the @xmath10 ( green , @xmath137 ) , and the new @xmath7 ( black , @xmath138 ) , are plotted against the true value of the exponent @xmath2 of the probability distribution samples are drawn from . clearly , below @xmath139 the @xmath10 estimator no longer works reliably . @xmath10 and @xmath7 work equally well in a range of @xmath140 . outside this range @xmath7 performs consistently better than the other methods . the inset shows the mean - square error @xmath141 of the estimated exponents . the ls - estimator has a much higher @xmath141 over the entire region , than the @xmath7-estimator . the blue dot represents the @xmath7 estimate for the zipf exponent of c. dickens `` a tale of two cities '' . clearly , this exponent could never reliably be obtained from the rank ordered distribution using @xmath10 , whereas @xmath7 works fine even for values of @xmath142 . [ fig : comparison ] , title="fig : " ] + controlling the fit region over which the power - law should be obtained therefore becomes a matter of restricting the sample space to a convenient @xmath143 . this can be used for dynamically controlling low - frequency cut - offs . these cut - offs are set to exclude states for which , @xmath144 where @xmath145 is the minimal number of times that any state - type @xmath21 is represented in the data set . this means that we re - estimate @xmath2 on @xmath143 with @xmath146 we see in eq . ( [ bayes4 ] ) that iteratively adapting @xmath60 to subsets @xmath147 , and then re - evaluating @xmath2 , requires to solve , @xmath148 where @xmath149 is the restricted sample - size and @xmath150 are the relative frequencies re - normalized for @xmath147 . @xmath151 is the index - set of @xmath147 . iterating this procedure either leads to a fixed point or to a limit cycle between two low - frequency cut - offs with two slightly different estimates for @xmath8 . these two possibilities need to be considered in order to implement an efficient stopping criterion for the iterative search of the desired low - frequency cut - off in the data . the algorithm therefore consists of two nested iterations . the `` outer iteration '' searches for the low - frequency cut - off , the `` inner iteration '' solves the implicit equation for the power - law exponent . the matlab code for the algorithm is found in @xcite , see appendix b for a tutorial . to test the proposed algorithm implementing the estimator @xmath7 , we first perform numerical experiments and then test its performance on a number of well known data sets . for 400 different values of @xmath2 , ranging from @xmath152 to @xmath153 , we sample @xmath132 data points @xmath154 , with @xmath155 states , with probabilities @xmath134 . we fit the data in three ways , using ( i ) least square fits ( ls ) , ( ii ) the csn algorithm @xmath10 providing estimates @xmath11 , and ( iii ) the implicit @xmath7 method providing estimates @xmath8 . in fig . [ fig : comparison ] we show these estimates for the power exponents , as a function of the true _ values _ of @xmath2 . the @xmath136 , @xmath10 , @xmath7 estimators are shown as the red , green , and black curves respectively . obviously @xmath7 and @xmath10 work equally well for power - law exponents @xmath2 with values @xmath140 . in this range the three approaches coincide . however , note that in the same region the mean square error , where @xmath156 is the number of repetitions , i.e. the number of data - sets we sampled from the @xmath157 , @xmath158 . @xmath159 is the value estimated for @xmath2 from the @xmath160th data set . depending on the estimator @xmath135 corresponds to @xmath11 ( @xmath10 ) , @xmath8 , ( @xmath7 ) , or the ls estimator . we used @xmath155 and @xmath161 for any given @xmath2 . ] @xmath141 for the ls method is much larger than for @xmath7 and @xmath10 . outside this range the assumptions and approximations used for @xmath10 start to lose their validity and both @xmath136 and @xmath7 estimates outperform the @xmath10 estimates . the inset also shows that @xmath7 consistently estimates @xmath2 much better than the @xmath136 estimator ( two orders of magnitude better in terms of @xmath141 ) for the entire range of @xmath2 . the blue dot in fig . [ fig : comparison ] represents the @xmath7 estimate for the zipf exponent of c. dickens ` a tale of two cities ' . clearly , this small exponent could never be obtained by @xmath10 , see also tab . [ table1 ] . .comparison of the estimators @xmath7 and @xmath10 on empirical data sets that were used in @xcite . these include the frequency of surnames , intensity of wars , populations of cities , earthquake intensity , numbers of religious followers , citations of scientific papers , counts of words , wealth of the forbes 500 firms , numbers of papers authored , solar flare intensity , terrorist attack severity , numbers of links to websites , and forest fire sizes . we added the word frequencies in the novel a tale of two cities " ( c. dickens ) . the second column states if @xmath46 or @xmath2 were estimated . the exponents reported in @xcite are found in column @xmath162 , those reproduced by us applying their algorithm to data @xcite is shown in column @xmath163 . the latter correspond well with the new @xmath7 algorithm . for values @xmath164 , @xmath165 can not be used . we list the corresponding values for kolmogorov - smirnov test for the two estimators , @xmath166 and @xmath167 . [ table1 ] [ cols="<,^,^,^,^,^,^",options="header " , ] we finally compare the new estimator @xmath7 on several empirical data sets that were used for demonstration in @xcite . in tab . [ table1 ] we collect the results . the second column states if @xmath2 or @xmath46 were estimated . column @xmath162 presents the value of the estimator @xmath10 as presented in @xcite . column @xmath163 contains the values of the same estimator using the data from @xcite and using the algorithm provided by @xcite . the results for the @xmath7 estimator agrees well with those of @xmath10 in the range where the latter works well . to demonstrate how @xmath7 works perfectly outside of the comfort zone of @xmath10 ( for @xmath164 ) , we add the result of the rank distribution of word counts in the novel a tale of two cities " ( charles dickens , 1859 ) , which shows an exponent of @xmath168 . this exponent can be fitted directly from the data using the proposed @xmath7 algorithm , while @xmath10 can not access this range , at least not without the detour of first producing a histogram from the data and then fitting the tail of the frequency distribution . the values for the corresponding kolmogorov - smirnov tests ( see e.g. @xcite ) for the two estimates , @xmath166 and @xmath167 , are similar for most cases . we discuss the generic problem of estimating power - law exponents from data sets . we list a series of questions that must be clarified before estimates can be performed . we present these questions in form of a decision tree that shows how the answers to those questions lead to different strategies for estimating power - law exponents . to follow this decision tree can be seen as a recipe for fitting power exponents from empirical data . the corresponding algorithms were presented and can be downloaded as matlab code . the two algorithms we provide are based on a very general ml estimator that maximizes an appropriately defined cross entropy . the method can be seen as a straight forward generalization of the idea developed in @xcite . the two estimators ( one for binned histograms and @xmath7 for raw data sets ) allow us to estimate power - law exponents in a much wider range than was previously possible . in particular , exponents lower than @xmath164 can now be reliably obtained . this work was supported in part by the austrian science foundation fwf under grant p29252 . b.l . is grateful for the support by the china scholarship council , file - number 201306230096 . 10 url # 1`#1`urlprefix[2]#2 [ 2][]#2 m.e.j . newman , _ power - laws , pareto distributions and zipf s law _ , contemporary physics 2005 ; * 46 * 32351 . m. mitzenmacher , _ a brief history of generative models for power - law and lognormal distributions _ , internet mathematics 2004 ; * 1 * 22651 kadanoff , et al . , _ static phenomena near critical points : theory and experiment _ , rev . 1967 ; * 39 * 395413 . d. sornette , _ critical phenomena in natural sciences _ , springer , berlin , 2006 . p. bak , c. tang , and k. wiesenfeld , _ self - organized criticality : an explanation of 1/f noise _ , phys . lett . 1987 ; * 59 * 38184 . simon , _ on a class of skew distribution functions _ , biometrika 1955 ; * 42 * 42540 . a.rka , and a.l . barabsi , _ statistical mechanics of complex networks _ , rev . mod . phys . 2002 ; * 74 * 4797 . barabsi , and a. rka , _ emergence of scaling in random networks _ , science 1999 ; * 286 * 509 - 12 . yule , _ a mathematical theory of evolution , based on the conclusions of dr . j. c. willis , f.r.s _ , phil . royal soc . b 1925 ; * 213 * 2187 . h. takayasu , a .- h . sato , and m. takayasu , _ stable infinite variance fluctuations in randomly amplified langevin systems _ , phys . 1997 ; * 79 * 96667 . c. tsallis , _ introduction to nonextensive statistical mechanics _ , springer , new york , 2009 . r. hanel , s. thurner , s , and m. gell - mann , _ how multiplicity of random processes determines entropy : derivation of the maximum entropy principle for complex systems _ , proc . usa 2014 ; * 111 * 690510 . b. corominas - murtra , r. hanel , and s. thurner , _ understanding scaling through history - dependent processes with collapsing sample space _ , usa 2015 ; * 112 * , 5348 - 53 . b. gutenberg , and c.f . richter , _ frequency of earthquakes in california _ , bull . 1944 ; * 34 * 18588 . k. christensen , l. danon , t. scanlon , and p. bak , _ unified scaling law for earthquakes _ proc . usa 2002 ; * 99 * 2509 - 13 . f. auerbach , _ das gesetz der bevlkerungskonzentration _ , petermanns geographische mitteilungen 1913 ; * 59 * 74 - 76 . x. gabaix , _ zipf s law for cities : an explanation _ , 1999 ; * 114 * 73967 . shaffer , _ spatial foraging in free ranging bearded sakis : traveling salesmen or lvy walkers ? _ , amer . j. primatology 2014 ; * 76 * 47284 . newman , and r.g . palmer , _ modeling extinction _ , oxford university press , 2003 . zipf , _ human behavior and the principle of least effort _ , addison - wesley , cambridge , massachusetts , 1949 . press , _ subjective and objective bayesian statistics : principles , models , and applications _ , wiley series in probability and statistics , 2010 . berger , _ statistical decision theory and bayesian analysis _ , springer , new york , 1985 . fisher , _ on an absolute criterion for fitting frequency curves _ , messenger of mathematics 1912 ; * 41 * 15560 . a. clauset , c.r . shalizi , and m.e.j . newman , _ power - law distributions in empirical data _ , siam review 2009 ; * 51 * 661703 . y. virkar , and a. clauset , _ power - law distributions in binned empirical data _ , annals of applied statistics 2014 ; * 8 * 89119 . a. deluca , and a. corral , _ fitting and goodness - of - fit test of non - truncated and truncated power - law distributions _ acta geophysica 2013 ; * 61 * 135194 a. broder , r. kumar , f. maghoul , p. raghavan , s. rajagopalan , r. stata , a. tomkins , and j. wiener , _ graph structure in the web _ , computer networks 2000 ; * 33 * 30920 . d.c . roberts , and d.l . turcotte , _ fractality and self - organized criticality of wars _ , fractals 1998 ; * 6 * 35157 . s. redner , _ how popular is your paper ? an empirical study of the citation distribution _ , epj b 1998 ; * 4 * 13134 . a. clauset , m. young , and k.s . gleditsch , _ on the frequency of severe terrorist events _ , journal of conflict resolution 2007 ; * 51 * 5887 . http://tuvalu.santafe.edu/@xmath169aaronc/powerlaws/ http://www.complex-systems.meduniwien.ac.at/ + si2016/r_plfit.m + http://www.complex-systems.meduniwien.ac.at/ + si2016/r_plhistfit.m h.s . heaps , _ information retrieval : computational and theoretical aspects _ , academic press , 1978 . g. herdan , _ type - token mathematics _ , gravenhage , mouton & co , 1960 . if events @xmath12 are drawn from a continuous sample space @xmath170 $ ] , for instance the magnitude of earthquakes , then the ` natural order ' of possible events is simply given by the magnitude @xmath12 of the observation . events @xmath12 are drawn from a continuous power - law distribution @xmath171 , with @xmath172)$ ] ( compare eq . ( [ norm ] ) first line ) . to work with well defined probabilities we have to bin the data first . probabilities to observe events within a particular bin depend on the margins of the @xmath19 bins @xmath173 , with @xmath174 and @xmath175 . the histogram @xmath26 counts the number @xmath24 of events @xmath12 falling into the bin @xmath176 , and the probability of observing @xmath12 in the @xmath21th bin is given by @xmath177 binning events sampled from a continuous distribution may have practical reasons . for instance data may be collected from measurements with different physical resolution levels , so that binning should be performed at the lowest resolution of data points included in the collection of samples . we will not discuss the ml estimator for binned data in detail but only remark that for given bin margins @xmath178 it is sufficient to insert @xmath179 of eq ( [ appa1 ] ) into eq . ( [ bayes3 ] ) with @xmath101 , to derive the appropriate ml condition for binned data . an algorithm for binned data ` r_plhistfit ` , where we assume the bin margins @xmath180 to be given , is found in @xcite . we point out that if margins for binning have not been specified prior to the experiments , then specifying the optimal margins for binning the data becomes a parameter estimation problem in itself , i.e. the optimal margins @xmath180 have to be estimated from the data as well . one major source of uncertainty in the estimates of @xmath2 from binned data is related to the uncertainty in choosing the upper and lower bounds @xmath90 and @xmath91 of the data , i.e. specifying the bounds of the underlying continuous sample space . binning becomes irrelevant for clean continuous data for the following reason . suppose we fix the sample space @xmath181 $ ] and cut this domain into @xmath109 bins of width @xmath182 . since the data @xmath183 is drawn from a continuous sample space , the chance for two observations @xmath184 and @xmath67 to be exactly equal becomes zero for @xmath185 , if @xmath109 has been chosen sufficiently large . then each bin almost certainly contains either one sample @xmath67 or none . the probability of observing @xmath12 then is asymptotically ( as @xmath186 approaches zero ) given by @xmath187 the parameter estimation problem of finding the optimal @xmath2 is equivalent to maximizing @xmath188 ( or equivalently @xmath189 ) with respect to @xmath2 . in this maximization problem @xmath186 becomes irrelevant and only the choice of @xmath190 and @xmath191 and the data @xmath12 remains relevant for the estimate . as a consequence , one obtains an equation @xmath192 for the ml estimate of the exponent @xmath2 over continuous sample spaces . equation ( [ bayes4 ] ) and eq . ( [ bayesx ] ) differ only in @xmath193 . in eq . ( [ bayes4 ] ) the normalization constant of discrete samples spaces gets used while in eq . ( [ bayesx ] ) @xmath193 is the normalization constant for a continuous sample space . switching between continuous and discrete sample spaces therefore is simply a matter of choosing the one or the other normalization constant in the algorithm . whether data should be assumed to be sampled from continuous or discrete sample spaces is not always totally clear . many measurements have an intrinsic resolution and implicitly bin the data . for instance if real numbers sampled in an experiment are given only with a three digit precision , such as @xmath194 and we know that @xmath195 and @xmath196 then we better treat the data as discrete data on @xmath197 if we have sufficiently many samples for the histogram over @xmath64 not to be flat . a primitive test to see whether one should regard data as sampled from a continuous sample space or not is to make a histogram over the unique values of the recorded data . if each distinct value appears only once in the data ( i.e. if the histogram over the unique data - points is flat ) then one should treat the sample - space as continuous . while for the discrete case we need not estimate @xmath90 and @xmath91 this remains necessary for the continuous case . the method of cutting the @xmath198 $ ] into segments of length @xmath186 and then taking @xmath186 to zero explains why typically tha _ primitive _ estimates , @xmath199 and @xmath200 , provides fairly good results . alternatively , strategies such as suggested in @xcite could be used to optimize the choices for @xmath90 and @xmath91 . however , this procedure can not be directly derived from bayesian arguments . neither will we discuss this approach in this paper nor implement such an option in ` r_plfit ` . however , bayesian estimates of @xmath90 and @xmath91 exist . although we will not discuss those estimators in detail here we will eventually implement them in ` r_plfit ` to replace the primitive estimates . the idea of constructing such estimators is the following . for instance , one asks how likely can the maximal value @xmath201 of the sampled data @xmath66 be found to be larger than some value @xmath82 . by deriving @xmath202)$ ] and @xmath203)$ ] , as a consequence , it becomes possible to derive bayesian estimators for @xmath90 and @xmath91 . the matlab function ` function out = r_plfit(data , varargin ) ` implements the algorithm discussed in the main paper . the function returns a struct ` out ` that contains information about the data , the data range , but most and for all ` out.exponent ` returns the estimated exponent of the power - law . whether the exponent ` out.exponent ` is the exponent @xmath2 of the sample distribution or the exponent @xmath46 of the frequency distribution of the data depends on how ` function out = r_plfit(data , varargin ) ` gets used as explained below . in the code the sample space @xmath60 is equivalent to a vector @xmath204 $ ] containing @xmath19 distinct event magnitudes @xmath47 , @xmath20 . + the variable ` data ` can be used to import data while a variable number of arguments can be set by ` varargin ` to tell the algorithm which type of data it should handle and to control the range of the data . by default the only argument that has to be set is ` data ` . ` r_plfit ` filters data from data points ` data<=0 , nan , inf ` . the data passed on to ` data ` can be * a vector of observations ` data ` @xmath205 $ ] ( default ) * a histogram ` data ` @xmath206 $ ] of recorded event types @xmath20 ` out = r_plfit(data , varargin ) ` can be used in three basic modes * ` out = r_plfit(x ) ` returns the estimated exponent @xmath2 of the * probability distribution * given the observation @xmath12 ( default ) * ` out = r_plfit(k,'hist ' ) ` returns the estimated exponent @xmath2 of the * probability distribution * given the histogram of observations @xmath73 * ` out = r_plfit(k ) ` returns the estimated exponent @xmath46 of the * frequency distribution * given the histogram of observations @xmath73 the third mode ` out = r_plfit(k ) ` is in fact identical to the first mode ` out = r_plfit(x ) ` , only that passing a histogram as sample data to the algorithm is identical to asking how many of the @xmath19 states @xmath21 have been observed @xmath44 times . but this is exactly the frequency distribution of the process , which possesses a tail with exponent @xmath207 . depending on the mode ` r_plfit ` returns the exponent @xmath2 or @xmath46 in ` out.exponent ` + * fitting with observations @xmath12 * : if we run ` out = r_plfit(x ) ` without further options ` r_plfit ` assumes by default that the data @xmath12 consists of natural numbers , and that the process samples have been sampled from the sample space @xmath208 , i.e. @xmath209 . if this is not the case one can either specify the data range using all @xmath19 _ unique _ values @xmath204 $ ] occurring in the data @xmath12 by using the option ` out = r_plfit(x,'urange ' ) ` . in order to define a fit range maximal and minimal data values taken into account can be set by ` out = r_plfit(x,'urange','rangemin',minval , ... ` ` ... ' rangemax',maxval ) ` such that ` r_plfit ` only takes into account data in the range ` minval ` @xmath210 ` maxval ` . to control the data range individually use ` out = r_plfit(x,'range',z ) ` . if the data has been sampled from a continuous sample space , and the histogram over the unique data is flat , i.e. each value in the data only appears once ( more or less ) , then one can tell ` r_plfit ` that the data is sampled from a continuous sample space by setting the option ` ' cdat ' ` , i.e. by running ` out = r_plfit(x,'cdat ' , ... ) ` . this option tells the algorithm to use the normalization constant for continuous sample spaces and estimates @xmath211 and @xmath212 . moreover , ` ' cdat ' ` implicitly sets the ` ' urange ' ` and the ` ' nolf ' ` option . ` ' nolf ' ` ( see below ) switches off the search of the algorithm for an optimal low frequency cut - off . + * fitting with histograms @xmath73 * : using histograms @xmath73 as input works in exactly the same way as for fitting @xmath12 if we want to estimate the exponent @xmath46 of the frequency distribution and use ` r_plfit ` in the ` out = r_plfit(k ) ` mode . if we use ` r_plfit ` in the ` out = r_plfit(k,'hist ' ) ` mode , the algorithm assumes by default that the sample space @xmath213 is given by @xmath214 $ ] . the option ` ' urange ' ` has no effect in this mode and gets ignored if set . otherwise one can again use the ` ' range ' ` property to set the event magnitudes @xmath213 ( the sample space ) using ` out = r_plfit(k,'hist','range',z ) ` . the ` ' minrange ' ` and ` ' maxrange ' ` options work in exactly the same way as before . + * dynamic low frequency cut - off * : by default ` r_plfit(data ) ` runs an iterative search for an optimal low frequency cut - off that is set at a range value @xmath47 such that the expected number of samples for @xmath47 equals the variable @xmath215 ( default value @xmath38 , reset using option ` ' nmin ' ` ) . this means the algorithm performs a low frequency cut - off for observations @xmath12 . if however ` maxval ` is smaller than the predicted cut - off then the low frequency cut - off has no effect . one should note that in the mode ` out = r_plfit(k ) ` the low frequency cut - off mechanism effectively acts as a high frequency cut - off with respect to the data @xmath12 . one can switch this mechanism off by setting the option ` ' nolf ' ` ( no low frequency cut - off ) . + the ` ' plot ' ` option , ` out = r_plfit(data , ... ` ` ... , ' plot ' ) ` , can be used for visualization . ` r_plfit ` plots the fit over the data in double logarithmic coordinates ( loglog plot ) . using the option ` ' figure ' ` behaves like ` ' plot ' ` but explicitly opens a new figure . ` ' exp_min ' ` can be used to specify the minimal search value for the exponents ( default is @xmath152 ) and ` ' exp_max ' ` to set the maximal search value ( default is @xmath216 ) . ` ' eps ' ` can be used to set the precision of the implicit algorithm ( default @xmath217 ) . several other options exist to control the performance of the algorithm , which all can be listed by using ` r_plfit('help ' ) ` in the command line , which prints a brief manual on the usage of ` r_plfit ` and available options . if one works with binned data , e.g. histogram data counting the number of events falling into exponentially scaled bins ( log - binning ) , then ` r_plhistfit ` needs to be used instead of ` r_plfit ` . the function ` function out = r_plhistfit(data , varargin ) ` like ` r_plfit ` , by default , uses only data as input and other variables can be set optionally . ` data ` is always a histogram @xmath73 that is a vector @xmath218 $ ] . bins can be specified by giving bin margins @xmath219 $ ] such thatevents counted in @xmath24 had a magnitude @xmath12 such that @xmath220 . usage , ` r_plhistfit(k,'margins',b ) ` . by default ` r_plhistfit ` assumes that @xmath221 . other options work similar to the ones available for ` r_plfit ` and can be reviewed by typing ` r_plhistfit('help ' ) ` in the matlab command line .	it has been repeatedly stated that maximum likelihood ( ml ) estimates of exponents of power - law distributions can only be reliably obtained for exponents smaller than minus one . the main argument that power laws are otherwise not normalizable , depends on the underlying sample space the data is drawn from , and is true only for sample spaces that are unbounded from above . here we show that power - laws obtained from bounded sample spaces ( as is the case for practically all data related problems ) are always free of such limitations and maximum likelihood estimates can be obtained for arbitrary powers without restrictions . here we first derive the appropriate ml estimator for arbitrary exponents of power - law distributions on bounded discrete sample spaces . we then show that an almost identical estimator also works perfectly for continuous data . we implemented this ml estimator and discuss its performance with previous attempts . we present a general recipe of how to use these estimators and present the associated computer codes .
in recent years the problem of an impurity coupled to a quantum bath has received a large attention in the field of ultracold atoms both theoretically and experimentally . in particular , the fermionic version of this problem , _ i.e. _ when the bath is a spin - polarized fermi sea , is already a well studied topic with different interesting perspectives addressing the attractive and repulsive branch as well as low dimensions @xcite . on the contrary , for the bosonic counterpart corresponding to an impurity immersed in a bose condensed medium , only very recently there have been experimental studies focusing on the excitation energy and spectral response of the polaron quasiparticle @xcite . previous experiments investigated mainly collisional and dissipation processes involving the bath @xcite . on the theoretical side , the bose polaron problem has been already addressed in a series of studies uitilizing different tools such as t - matrix @xcite and perturbation @xcite approaches , variational wavefunction @xcite as well as quantum monte carlo ( qmc ) methods @xcite . low dimensions , and in particular one - dimensional ( 1d ) configurations , enrich the bose polaron problem with some peculiar features that are worth investigating . first of all the enhanced role of quantum fluctuations capable to destroy the off - diagonal long - range order responsible for bose - einstein condensation even at zero temperature . secondly , the possibility to achieve strongly correlated regimes for the quantum bath , where the bosons approach the so called tonks - girardeau ( tg ) limit of fermion - like impenetrable particles @xcite . the use of confinement induced resonances of the @xmath0-wave scattering amplitude represents for these 1d systems a powerful tool , allowing for a wide tunability of both the interactions within the bath and between the bath and the impurity @xcite . for example , in ref . @xcite a 1d mixture of k impurities in a gas of rb atoms was realized and different values of the impurity - bath interaction , obtained by varying the magnetic field , were probed . the gas parameter in the bath can also be tuned by changing the density or the effective mass of the atoms by means of a lattice potential @xcite . in this work we investigate theoretically the bose polaron problem when the quantum bath is modeled by a lieb - liniger gas with contact interactions @xcite and the impurity is coupled to the bath via a different @xmath1-like potential that can be both attractive and repulsive . different regimes of the surrounding medium are considered and analyzed as a function of the impurity - boson coupling strength : i ) the bath is weakly interacting and well described by the mean - field theory , ii ) interactions in the bath are relatively strong and beyond mean - field effects are important , and finally iii ) the bath is in the tg regime and some analytical results are available thanks to the bose - fermi mapping @xcite . we characterize the bose polaron by calculating its binding energy , effective mass , and contact parameter . furthermore , we consider both the case of a mobile impurity having the same mass of the particles in the medium and the case of a static impurity , corresponding to the limit of a much larger impurity to boson mass ratio . the results show that the polaron energy reaches a constant value for large repulsive impurity - boson interaction , whereas in the opposite limit of large attraction the impurity gets deeply bound to the bath . unless the bosons are impenetrable particles , this binding energy is found to be much larger than the one of the dimer state , the solution to the two - body problem in vacuum . the tendency of bosons attracted by the impurity to form large clusters involving many particles is also evident from the density profile of the medium around the impurity . the size of these clusters , as well as their binding energy , increases with decreasing interaction strength within the bath , indicating an instability of the weakly interacting lieb - liniger gas towards collapse around an attractive impurity . the effective mass of the polaron moving in a weakly interacting medium exhibits a sharp increase as a function of the coupling , both on the repulsive and on the attractive side of the impurity - bath interaction . already for a moderate coupling strength the impurity gets very heavy , thus realizing the so - called `` self - localization '' regime predicted by landau and pekar for polarons in crystals @xcite . this picture was also proposed for bose polarons in three dimensions ( 3d ) on the basis of a frhlich - type model @xcite which , however , does not capture the relevant physics at strong coupling @xcite . on approaching the tg limit , instead , the effective mass increases sizeably only for large values of the coupling on the repulsive side and saturates at twice the bare mass on the attractive side . it is worth pointing out that the results mentioned above are directly relevant for experiments on few impurities in a quantum bath where the binding energy can be measured using spectroscopic techniques @xcite and the effective mass from the study of collective modes in harmonic traps @xcite . the paper is organized as follows : in sec . ii we introduce the hamiltonian of the system and we review or derive some analytical results that are obtained using perturbation theory in the limit of weak interactions or the bose - fermi mapping in the tg limit . in sec . iii we briefly provide some details about the qmc techniques used to address this specific problem and about the physical observables calculated in the simulations . iv contains a detailed discussion of the results and a comparison with the analytical predictions available for specific regimes of parameters . finally , in sec . v we draw our conclusions and we outline possible future prospects of this work . we consider the following hamiltonian @xmath2 describing a 1d system of @xmath3 identical bosons with mass @xmath4 interacting via a repulsive contact potential of strength @xmath5 . a single impurity of mass @xmath6 is coupled to the particles of the bath via another contact potential characterized by the strength @xmath7 . the coordinates of the bosons and the impurity are denoted respectively by @xmath8 ( @xmath9 ) and @xmath10 . we introduce the dimensionless parameters @xmath11 where @xmath12 is the density of the bath in the 1d box of size @xmath13 . the first parameter gives the strength of interactions within the bath which can range from the weakly correlated mean - field regime ( @xmath14 ) to the strongly correlated tg regime ( @xmath15 ) , where bosons are impenetrable and behave similarly to a gas of spinless fermions . the parameter @xmath16 is instead related to the coupling between the impurity and the bath and , contrary to @xmath17 , can be either positive or negative depending on the sign of @xmath7 . notice that for equal masses ( @xmath18 ) and for equal coupling strengths ( @xmath19 ) the equality @xmath20 holds . the above hamiltonian ( [ hamiltonian ] ) of the clean bath without the impurity was solved exactly by lieb and liniger @xcite for a system in the thermodynamic limit and with periodic boundary conditions . the ground - state energy is found as a function of the parameter @xmath17 in the form @xmath21 where the dimensionless energy per particle @xmath22 is obtained by solving a pair of coupled integral equations @xcite . the scale of energy in eq . ( [ liebliniger ] ) is chosen as @xmath23 , with @xmath24 , corresponding to the fermi energy of a gas of spinless fermions with the same mass and density . some limits of @xmath22 can be derived analytically : if @xmath14 , one finds @xmath25 ; in the opposite limit , @xmath15 , one has instead @xmath26 . as we already mentioned , the former result holds in the weak - coupling regime and includes the bogoliubov term and the first beyond mean - field correction whereas , in the latter , the leading term corresponds to the energy of the equivalent non - interacting fermi gas . the binding energy of the impurity is defined as the energy difference between the ground state of the system with and without the impurity . the former can be written as @xmath27 \ ; , \label{groundstate}\ ] ] allowing one to express the energy difference @xmath28 in terms of the dimensionless function @xmath29 which yields the polaron energy in units of the scale @xmath30 . in the limit of small @xmath17 and small @xmath16 one can determine the energy and the effective mass of the polaron by using the perturbation approach based on the bogolubov approximation as it is outlined in ref . @xcite . by introducing the mass ratio @xmath31 of the bosonic to the impurity mass , the result for the polaron energy reads : @xmath32 and for the effective mass one finds @xmath33 the first term in eq . [ enerpert ] is the mean - field contribution proportional to @xmath7 , whereas next - to - leading order corrections to both @xmath34 and the effective mass are proportional to @xmath35 and are independent of the sign of the impurity - boson interaction . we also notice that in the limit @xmath36 , for a fixed value of @xmath16 , both these corrections diverge , signalling an instability of the medium surrounding the impurity when boson - boson interactions are suppressed . the situation is different in the 3d case where , in the same limit , only the effective mass exhibits a divergent behavior @xcite . we consider explicitly two particular values of the mass ratio @xmath37 : the case of equal masses ( @xmath38 ) and the case of a static impurity with infinite mass ( @xmath39 ) . for the former case , eqs . ( [ enerpert ] ) and ( [ meffpert ] ) give @xmath40 where we have set @xmath41 . the binding energy of a static impurity , instead , is found to be : @xmath42 these two cases are of special interest because an exact solution is available when the bath is in the tg limit ( @xmath43 ) . the binding energy and the effective mass of the equal mass case ( @xmath38 ) were calculated by mcguire @xcite with the result @xmath44 \label{mcguire1}\ ] ] and the two results @xmath45 holding , respectively , for positive and negative values of the coupling constant @xmath7 . some limits of the above eqs . ( [ mcguire1])-([mcguire2 - 2 ] ) are worth discussing : i ) if @xmath46 one finds @xmath47 , yielding a polaron energy equal to the chemical potential of the surrounding fermi gas , whereas the effective mass diverges as @xmath48 . ii ) if @xmath49 , then @xmath50 and @xmath51 , corresponding to the binding energy and the mass of a dimer in vacuum . in the case of a tg gas with a static impurity ( @xmath39 ) one proceeds by considering the impurity in the center of a large box of size @xmath13 with impenetrable walls and by calculating the phase shift of each single - particle state generated by the impurity contact potential with strength @xmath7 . the ground - state energy difference @xmath28 is readily calculated yielding the result @xmath52 \nonumber\\ & -&[1-\theta(\eta)]\frac{\eta^2}{4\pi^2 } \;. \label{mustatic}\end{aligned}\ ] ] the term involving the heaviside function @xmath53 , where @xmath54 if @xmath55 and zero otherwise , accounts for the binding energy of the dimer when @xmath56 . also in this case we can easily extract the following limiting behaviors : i ) if @xmath46 , then @xmath57 and ii ) if @xmath49 one finds the result @xmath58 . notice that the energy of the mobile impurity ( [ mcguire1 ] ) in the limit of infinite repulsion is twice the corresponding energy of the static impurity ( [ mustatic ] ) , even though the effective mass ( [ mcguire2 - 1 ] ) diverges in the same limit . this is due to the kinetic energy contribution of the mobile impurity within the region of space delimited by the two nearest neighbour particles of the bath acting as impenetrable barriers . another asymptotically exact result for a static impurity ( @xmath39 ) is obtained when interactions within the bath are weak ( @xmath14 ) and there is a strong impurity - boson repulsion ( @xmath46 ) . in this case , the polaron energy coincides with the excitation energy of a dark soliton @xcite @xmath59 this excited state of the gas is indeed stationary and is characterized by a zero in the density profile . an important quantity describing the interaction between the impurity and the bath is the contact @xmath60 , defined as the value of the boson density at the impurity position normalized by the bulk density : @xmath61 . by using the hellmann - feynman theorem @xcite one can relate the value of @xmath60 to the derivative of the equation of state with respect to the impurity - boson coupling constant @xmath62 in the case of equal masses ( @xmath38 ) the contact parameter can be derived analytically in the weak - coupling limit , where eq . ( [ enerpert1 ] ) yields @xmath63 , and in the tg limit , where from eq . ( [ mcguire1 ] ) one finds @xmath64 from the above result we see that @xmath65 when @xmath16 is large and positive and @xmath66 in the opposite limit of large and negative values of @xmath16 . the calculations are performed using qmc numerical techniques . a thorough description of these methods can be found elsewhere @xcite . here we provide a detailed explanation of the choice of the wave function used as a trial many - body state in the variational monte carlo ( vmc ) and for importance sampling in the diffusion monte carlo ( dmc ) scheme . a similar scheme was used in the recent study of an impurity immersed in a 3d bose - einstein condensate @xcite . we consider a system of @xmath3 identical particles and one impurity contained in a 1d box of size @xmath13 with periodic boundary conditions . the trial wave function is comprised of two terms : the first takes into account correlations among the particles in the bath and the second between these particles and the impurity . both pair correlation functions satisfy boundary conditions when the two particles meet to enforce the contact interactions present in the hamiltonian ( [ hamiltonian ] ) . the trial wave function is written as @xmath67 where the multidimensional vector @xmath68 denotes the coordinates of both the @xmath3 particles of the bath and the impurity . the boson - boson correlation factor is chosen as the solution of the two - body problem @xmath69 if @xmath70 and @xmath71 otherwise . the value of @xmath72 is fixed by the condition @xmath73 ensuring that both @xmath74 and its first derivative are continuous at the matching point @xmath75 . the phase shift results from the bethe - peierls contact condition imposed by the interatomic potential and is given by @xmath76 . for the impurity - boson correlation term one should distinguish between the case of repulsive ( @xmath77 ) and attractive ( @xmath56 ) interactions . the former case can be dealt with in a way similar to the boson - boson interaction and one is lead to write @xmath78 when @xmath79 , while @xmath80 if @xmath81 . here @xmath72 is chosen such that @xmath82 in terms of the matching point @xmath83 and the phase shift is given by @xmath84 . in the case of attractive interactions a two - body bound state exists between the impurity and a particle of the bath and we choose the correlation term as a linear combination of the solutions corresponding to a negative two - body energy @xmath85 when @xmath79 and @xmath80 otherwise . the value of @xmath72 depends on the matching point @xmath83 and is determined through the contact condition imposed by the @xmath1-potential of strength @xmath7 : @xmath86 . the two matching points @xmath75 and @xmath83 , entering respectively the boson - boson and the impurity - boson correlation term , are variational parameters optimized by minimizing the expectation value of the hamiltonian on the many - body state @xmath87 . calculations are carried out by separately computing the ground - state energy of the system with @xmath3 bosons and one impurity , @xmath88 , and of the clean system with @xmath3 bosons only , @xmath89 . the polaron energy is determined as the difference @xmath90 , in units of the fermi energy . the effective mass is calculated by following the evolution in imaginary time of the impurity @xmath91 , being @xmath92 , and by extracting the diffusion constant @xmath93 where @xmath94 is the diffusion constant of a free particle @xcite . the total atom numbers @xmath3 range from @xmath95 to @xmath96 in order to control finite - size effects and extrapolate the results to the thermodynamic limit . the density profiles @xmath97 of the bosons in the vicinity of the impurity are calculated by making use of the standard extrapolation technique @xcite @xmath98 where @xmath99 is the `` mixed '' estimator of the density , obtained as the output of a dmc simulation , and @xmath100 is the `` variational '' estimator obtained from the corresponding vmc calculation . the contact parameter @xmath60 is extracted directly from the value of @xmath97 at the position of the impurity . we first discuss the results on the binding energy for the mobile ( @xmath38 ) and the static ( @xmath39 ) impurity . in fig . [ fig1 ] we show the energy of the mobile impurity as a function of the coupling strength @xmath16 , ranging from large negative to large positive values , when the interaction parameter @xmath17 in the bath is kept fixed . for this latter we consider values varying from the weakly coupled regime , @xmath14 , to the tg regime where @xmath43 . the corresponding results for the static impurity are presented in fig [ fig2 ] . notice that for @xmath101 we subtract from the binding energy @xmath34 the contribution from the two - body bound state given by @xmath102 . the exact polaron energies corresponding to the tg regime and given by eqs . ( [ mcguire1 ] ) and ( [ mustatic ] ) are shown as solid lines in fig . [ fig1 ] and fig . [ fig2 ] , respectively for @xmath38 and for @xmath39 . as an important benchmark test we find that the values of @xmath34 obtained from our qmc simulations with @xmath43 perfectly reproduce these results . in the regime of small values of @xmath103 , our results also recover the expansions from perturbation theory given by eqs . ( [ enerpert1 ] ) and ( [ enerpert0 ] ) . in particular , both for @xmath38 and for @xmath39 , we find that the range of values of @xmath103 where the perturbation expansion agrees well with the calculated polaron energy becomes larger as @xmath17 increases . in fact we notice that , when @xmath17 is large , from eq . ( [ enerpert ] ) one finds @xmath104 . remarkably this result agrees with the expansion of eq . ( [ mcguire1 ] ) and ( [ mustatic ] ) , respectively for the mobile and static impurity , up to values of the impurity - boson coupling constant on the order of @xmath105 . for smaller values of @xmath17 , typically @xmath106 , the applicability of perturbation theory is instead limited to the region where @xmath107 . if @xmath16 is large and positive the energy of both the mobile and the static impurity tends to saturate to a value that becomes smaller with decreasing @xmath17 [ see panel ( b ) of fig . [ fig1 ] and fig . [ fig2 ] ] . in the tg regime ( @xmath43 ) this asymptotic energy coincides with the energy @xmath30 of adding an extra particle to the bath in the mobile case and with @xmath108 in the static case . as already discussed in sec . ii , this energy difference arises from the kinetic energy contribution of the mobile impurity . a similar difference persists also for smaller values of @xmath17 : for example , at @xmath109 and @xmath110 , we find @xmath111 and @xmath112 respectively for the @xmath38 and @xmath39 case . we notice that , in this latter limit of large @xmath16 and small @xmath17 , the energy of the static impurity is expected to coincide with the excitation energy ( [ soliton ] ) of a dark soliton as determined using the mean - field gross - pitaevskii equation . indeed , in panel b ) of fig . [ fig2 ] , good agreement between the two energies is found for @xmath109 and also for @xmath113 . in the opposite regime of large and negative values of @xmath16 , the energy difference @xmath114 tends , when @xmath43 , respectively to @xmath115 and to @xmath116 in the mobile and in the static case . here , the impurity forms a two - body bound state with one of the particles of the medium which is then missing from the fermi sea of the tg gas . notice that , similarly to the case of @xmath117 , the binding energy @xmath34 of the mobile impurity is larger by a factor of two compared to the energy of the static one . for smaller values of @xmath17 our results indicate that @xmath114 is always negative and grows unbounded as @xmath49 [ see panel ( a ) of fig . [ fig1 ] and fig . [ fig2 ] ] . this behavior arises because , for attractive interactions between the impurity and the bath and for not too strong repulsion within the bath , many particles of the medium tend to cluster around the impurity producing a large negative binding energy for the polaron . an interesting question concerns the value of the polaron energy when @xmath16 is fixed and the interaction strength within the bath gets weaker and weaker ( @xmath36 ) . at the level of perturbation theory [ see eq . ( [ enerpert ] ) ] the answer to this question is that @xmath34 becomes large and negative irrespective of the sign of @xmath16 . this result differs from the corresponding situation in 3d where perturbation theory predicts that , when interactions within the bath vanish , the polaron energy reduces to the mean - field value , proportional to the interspecies coupling constant @xcite . for positive values of @xmath16 , a proper answer to the question , going beyond the result of perturbation theory , is provided by panel ( b ) of fig . [ fig1 ] and fig . [ fig2 ] : when @xmath118 the polaron energy decreases to zero and for the static impurity the behavior of @xmath34 is correctly described by the dark - soliton excitation energy given in eq . ( [ soliton ] ) . in the case of @xmath101 , panel ( a ) of fig . [ fig1 ] and fig . [ fig2 ] shows that the binding energy of the impurity grows large and negative as @xmath17 decreases for a fixed value of @xmath16 . however , in this limit , one can expect that either i ) @xmath119 , indicating the instability of the non - interacting bath in the presence of the impurity , or ii ) @xmath34 saturates to a finite energy , indicating that even a tiny repulsion in the bath is enough to stabilize the polaron . we notice that the value of @xmath29 refers to the polaron energy in the thermodynamic limit and that the @xmath36 limit is intended to be taken after the one of @xmath120 . of course , in the opposite case of a strictly non - interacting bath with a finite number of particles , the polaron energy would trivially diverge when the number @xmath3 increases . we address the question for the mobile impurity with mass ratio @xmath38 , when the interaction between the impurity and the bath is attractive and kept fixed at the value @xmath121 . the results of the inverse energy @xmath122 are shown in fig . [ fig3 ] as a function of decreasing values of @xmath17 . we find that @xmath122 decreases in absolute value as @xmath17 decreases , even though the result ( [ enerpert1 ] ) of perturbation theory fails completely in describing the trend of the calculated binding energies . a simple linear fit to the data extrapolates to a value compatible with @xmath123 when @xmath124 , given that error bars are significantly large . our findings are thus compatible with the above case i ) , showing an instability of the weakly repulsive bath towards a collapse around the impurity . a similar behavior is expected for the static impurity with @xmath39 . this latter result is in contrast with the binding energy of a static impurity in 3d and resonantly interacting with the medium which was found to approach a finite value in the limit of a vanishing repulsion within the bath @xcite . as a function of the parameter @xmath17 and for the fixed value @xmath121 of the impurity - boson coupling constant . the line is a linear fit to the data and the shadow region shows the statistical uncertainty of the fit . the dashed line is the prediction from perturbation theory given in eq . ( [ enerpert1]).,width=264 ] as a function of the impurity - boson interaction parameter @xmath16 and for different values of the coupling strength @xmath17 within the bath . both values corresponding to @xmath16 positive and negative are shown in the same graph . the @xmath43 curve corresponds to the exact results by mcguire of eqs . ( [ mcguire2 - 1 ] ) and ( [ mcguire2 - 2 ] ) in the tg limit . dotted lines refer to the result of perturbation theory given in eq . ( [ meffpert1]).,width=264 ] in fig . [ fig4 ] we show the results of the polaron effective mass in the case of the mobile impurity with mass ratio @xmath38 . we find that , for a given value of @xmath17 , the inverse effective mass decreases as @xmath103 increases both for repulsive and attractive interactions . we also notice that , in the tg limit of @xmath43 , we recover the exact results obtained by mcguire and given in eqs . ( [ mcguire2 - 1 ] ) and ( [ mcguire2 - 2 ] ) . furthermore , the comparison with the prediction ( [ meffpert ] ) of perturbation theory shows that , similarly to the case of the energy @xmath34 , the range of values of @xmath103 where agreement is found gets larger as @xmath17 increases . on the attractive side of impurity - bath interactions , the exact tg - gas result in eq . ( [ mcguire2 - 2 ] ) yields @xmath125 in the limit of @xmath16 large and negative . for the strongly interacting medium with @xmath126 we find in the same limit that the effective mass seems to saturate to @xmath127 . one should stress here that for large attractions the calculation of the effective mass requires increasingly longer simulation times resulting in larger error bars . for smaller values of @xmath17 , both on the attractive and on the repulsive side , the value of @xmath128 becomes very large for @xmath129 , if @xmath113 , and already for @xmath130 if @xmath109 . it is worth noticing that this rapid increase of the effective mass as a function of the impurity - bath coupling does not occur in the 3d counterpart of the bose polaron problem , where in the limit of resonant interaction between the impurity and the bath one finds @xmath131 ( see ref . @xcite ) . an increase of the effective mass as a function of the impurity - bath coupling has been reported in the experiment of ref . @xcite both for attractive and repulsive interactions . in this experiment a cloud of k impurities immersed in a 1d bath of rb atoms is suddenly released after compression with an optical potential and the rate of increase of its axial size is measured for different fixed values of the interaction strength between the impurities and the bath . the connection with the effective mass of the impurities is provided by interpreting the normalized width of the cloud with @xmath132 . the coupling constant of the bath was @xmath133 and values of @xmath16 as large as @xmath134 were produced , resulting in a maximum measured decrease of the normalized width by a factor of roughly 0.6 . . distances are in units of the inverse 1d density @xmath135 . the different curves correspond to different values of the parameter @xmath16 characterizing the strength of impurity - bath repulsive interactions . the various panels refer to different values of the coupling constant within the bath : panel ( a ) @xmath43 , panel ( b ) @xmath136 and panel ( c ) @xmath109.,width=302 ] . distances are in units of the inverse 1d density @xmath135 . the different curves correspond to different values of the parameter @xmath103 characterizing the strength of impurity - bath attractive interactions . the various panels refer to different values of the coupling constant within the bath : panel ( a ) @xmath43 , panel ( b ) @xmath136 and panel ( c ) @xmath109.,width=302 ] in fig . [ fig5 ] and fig . [ fig6 ] we show the density profiles of the bath as a function of the distance from the mobile impurity with mass ratio @xmath38 and for different values of both the impurity - boson and the boson - boson coupling constant . [ fig5 ] refers to repulsive interactions between the impurity and the bath , whereas fig . [ fig6 ] refers to attractive interactions . by increasing the value of @xmath16 in the case of repulsive interactions , the density of bosons in the close vicinity of the impurity decreases until a hole , completely empty of particles , is created for very large @xmath16 . the size of the hole strongly depends on the interaction parameter within the bath : it is on the order of the interparticle distance for the largest value ( @xmath43 ) and it extends to up to @xmath137 interparticle distances for the smallest one ( @xmath109 ) [ see panels ( a)-(c ) in fig . [ fig5 ] ] . we also notice that , for @xmath43 and for the largest value of @xmath16 , friedel - type oscillations , typical of the fermionic nature of the tg gas , are visible in the density profile . the results of fig . [ fig6 ] , instead , feature a peak of the boson density around the position of the impurity which becomes higher as the strength of the attraction increases . also in this case the size of the peak gets larger as @xmath17 decreases . for the largest @xmath17 [ see panel ( a ) in fig . [ fig6 ] ] the size of the peak is smaller than the interparticle distance and at large values of @xmath103 only one particle of the bath is , on average , close to the impurity forming a bound dimer with energy @xmath138 . indeed , as already mentioned when discussing panel ( a ) of fig . [ fig1 ] , the energy of a polaron in a tg gas tends to @xmath138 when @xmath16 is large and negative . on the contrary , as the value of @xmath17 decreases , the density peak becomes wider and involves more and more particles of the bath [ see panels ( b ) and ( c ) in fig . [ fig6 ] ] . as a consequence , the binding energy of the impurity increases in absolute value and one is approaching the situation of an unstable weakly interacting gas as shown in fig . [ fig3 ] . finally , in fig [ fig7 ] , we show the results on the contact parameter @xmath60 defined in eq . ( [ contact ] ) . as explained in the previous section , the value of @xmath60 is determined from the boson density @xmath97 at the impurity position normalized by the bulk density @xmath139 . we notice that we reproduce the exact result of eq . ( [ contacttg ] ) in the tg regime . furthermore , as compared to the @xmath43 case , we see that for smaller values of @xmath17 the contact parameter drops faster with increasing positive @xmath16 and diverges faster with increasing negative @xmath16 . qualitatively similar results for the density profiles and the contact parameter are obtained in the case of the static impurity with mass ratio @xmath39 . of a mobile impurity with mass ratio @xmath38 as a function of the impurity - boson interaction parameter @xmath16 and for different values of the coupling strength @xmath17 within the bath . both values corresponding to @xmath16 positive and negative are shown in the same graph . the solid line corresponds to the exact result ( [ contacttg ] ) holding in the tg limit . dotted lines refer to the perturbative result holding for small values of @xmath16.,width=283 ] by using exact " qmc numerical methods we investigated the properties of a bose polaron in 1d as a function of both the coupling strength between the impurity and the bath and within the bath . for a given impurity - bath interaction strength we find that the repulsive polaron can never exceed the energy reached when the bath is in the tg regime . on the contrary , the binding energy of the attractive polaron lies always below the energy of a dimer in vacuum and becomes increasingly large as the repulsion within the medium is reduced , thereby signalling an instability of the weakly interacting gas towards collapse around the impurity position . furthermore , in the regime of a weakly repulsive medium , the polaron effective mass is found to increase sharply with the strength of the impurity - bath coupling . such a heavy impurity , practically immobile within the medium , realizes the long - sought after regime of self - localization " of the strongly coupled landau - pekar polaron . interesting future prospects of this work include the study of a finite concentration of impurities in a 1d quantum bath . this topic naturally leads to the issue of interaction between impurities , phase separation in the two - component gas and the properties of magnetic excitations in the miscible mixture . useful discussions with g.e . astrakharchik are gratefully acknowledged . this work was supported by the quic grant of the horizon2020 fet program and by provincia autonoma di trento .	we present a theoretical study based upon quantum monte carlo methods of the bose polaron in one - dimensional systems with contact interactions . in this instance of the problem of a single impurity immersed in a quantum bath , the medium is a lieb - liniger gas of bosons ranging from the weakly interacting to the tonks - girardeau regime , whereas the impurity is coupled to the bath via a different contact potential producing both repulsive and attractive interactions . both the case of a mobile impurity , having the same mass as the particles in the medium , and of a static impurity with infinite mass are considered . we make use of exact numerical techniques that allow us to calculate the ground - state energy of the impurity , its effective mass as well as the contact parameter between the impurity and the bath . these quantities are investigated as a function of the strength of interactions between the impurity and the bath and within the bath . in particular , we find that the effective mass rapidly increases to very large values when the impurity gets strongly coupled to an otherwise weakly repulsive bath . this heavy impurity hardly moves within the medium , thereby realizing the `` self - localization '' regime of the landau - pekar polaron . furthermore , we compare our results with predictions of perturbation theory valid for weak interactions and with exact solutions available when the bosons in the medium behave as impenetrable particles .
tensor computations play a crucial role in the studies of general relativity ( gr ) . however , in typical cases those calculations are tediously complicated . fortunately , with the development of symbolic calculation in computer technology , a number of computer packages have been built to calculate the gr tensors , including grtensor @xcite , xact @xcite , ricci @xcite , and so on . general relativitists would most likely find whatever functionality they need in those packages and work with them . however , in case the researchers do have a need to modify some internals of the packages , or hope to have a complete understanding of the package they use , things become complicated . although many packages are kindly provided open source , they have in general a heavy code base . for example , the latest xact package has over 20,000 lines of code ( loc ) , where the xtensor.m file alone has more than 8,000 loc . as another example , the ricci package has over 7,000 loc . there is definitely nothing wrong with heavy weight packages which provide as many functionalities as possible . but on the other hand , it would also be helpful to have lightweight packages which are easy for the users to understand the underlying mechanisms and modify them when needed . the _ mathgr _ package is built for this purpose . currently , _ mathgr _ has about 500 loc , and those numbers will not grow significantly in the future . the package is also built modularly , separating the tensor manipulation , gr definitions , integration by parts , specific model definition , general utilities , display parser into different files . while keeping to be lightweight , _ mathgr _ still provides competing functionalities for gr computations , with fast speed . the functionalities of _ mathgr _ include * tensor simplifications with symmetries . symmetries , anti - symmetries for any subset of indices , and any permutational symmetries defined by the _ mathematica _ cycles can be brought into unique forms , such that if there can be cancellation between terms in an expression , they indeed cancel each other after simplification . * tensor calculations with either abstract or explicit ( or mixed ) indices , and decomposition from abstract indices to explicit ones . for example , tensors with abstract tensor indices such as the ones inside the calculation of the ricci scalar ( where the dummy indices denotes summation throughout this paper and assumed by the package ) @xmath0 can be calculated and simplified , and later decomposed if needed , with builtin command into @xmath1 where dot denotes derivative with respect to the 0 index , and @xmath2 run from @xmath3 to @xmath4 and @xmath5 run from @xmath6 to @xmath4 . alternatively , one can also decompose the @xmath2 indices into completely explicit indices @xmath7 , or decompose @xmath5 indices into explicit indices @xmath8 . * simplification with total derivatives . when expanding an action of a physical system , total derivatives can either be dropped or reduced into boundary terms . _ mathgr _ can try to reduce a given expression into total derivatives plus the rest of the terms which are minimal under various conditions . , or the remaining terms has minimal leaf count . however , in the leaf count case , _ mathgr _ can not guarantee to give the simplest result because there is no general algorithm for such reduction as far as we know . ] for example , @xmath9\end{aligned}\ ] ] * cosmic perturbations . there is a builtin model calculating gr tensors for the frw metric with adm type perturbations . this model can on the one hand be used directly for the research of inflationary cosmology , and on the other hand provide an example for writing models . * make use of _ mathematica _ s builtin tensor engine . since _ mathematica _ 9.0 , a symbolic tensor engine is introduced . this engine is powerful and fast in reducing tensors with symmetries into unique forms . however , the interface of the tensor engine is in the coordinate independent form . the input looks like : @xmath10 arrays[{dim , dim , dim } , complexes , antisymmetric[{1 , 2 } ] ] ) ; tensorreduce [ tensorcontract[t\[tensorproduct]t , { { 1 , 5 } , { 2 , 6 } , { 3 , 4 } } ] - tensorcontract[t\[tensorproduct]t , { { 1 , 4 } , { 2 , 6 } , { 3 , 5 } } ] ] } } & & \verb!$assumptions = ( t ! \in \verb!arrays[{dim , dim , dim } , complexes , antisymmetric[{1 , 2 } ] ] ) ; ! \\ \nonumber & & \verb!tensorreduce[tensorcontract[t ! \otimes \verb!t , { { 1 , 5 } , { 2 , 6 } , { 3 , 4 } } ] ] ! \\ \nonumber & & \quad - \verb!tensorreduce[tensorcontract[t ! \otimes \verb!t , { { 1 , 4 } , { 2 , 6 } , { 3 , 5 } } ] ] ! \endaccsupp{}\end{aligned}\ ] ] and the corresponding output is @xmath11 = -2 tensorcontract[t ! \otimes \verb!t , { { 1 , 4 } , { 2 , 6 } , { 3 , 5}}]!\end{aligned}\ ] ] + for many physicists , the above notation is not obvious , especially in case those notations come together with an expression with thousands of terms . _ mathgr _ makes use of the above engine , but the input and output now takes the form + @xmath12 t[dn@"c " , dn@"a " , dn@"b " ] - t[dn@"a " , dn@"b " , dn@"c " ] t[dn@"a " , dn@"c " , dn@"b " ] ] ( * -2 t[dn["a " ] , dn["b " ] , dn["c " ] ] t[dn["a " ] , dn["c " ] , dn["b " ] ] * ) } } \mathrm{simp}[t_{abc}t_{cab } - t_{abc}t_{acb } ] \qquad\rightarrow\qquad - 2 t_{abc}t_{acb } \endaccsupp{}\end{aligned}\ ] ] + on the other hand , for people using _ version lower than 9.0 , _ mathgr _ provides a simple tensor simplification engine , which can still do some calculations such as cosmic perturbations , but is not powerful enough to simplify tensor with symmetries ( except totally symmetric ) or permutation of indices . with the lightweight code base and modularity , we hope _ mathgr _ is helpful for researchers who have a demand to modify the internal functionality of a package , or those who want to take part of the internal technology from the package to build their own tools . moreover , the development is test driven . unit tests and integrated tests are provided thus the modifications by users are safer if all tests are passed . _ mathgr _ can be installed with the following steps : * download the _ mathgr _ package from https://github.com/tririver/mathgr/releases/tag/v0.1 . ( future updates will also be hosted on https://github.com/tririver/mathgr . ) * unpack the package . * rename the directory name ` mathgr-0.1 ` ( or any other version number ) to ` mathgr ` . note that ` mathgr ` is the directory which contains ` mathgr.m ` together with other files . * setup path so that _ mathematica _ can find and load _ mathgr_. two options are provided here : * * install : move the ` mathgr ` directory into a directory that _ mathematica _ can find by default . one can find out the default directory by evaluating ` $ path ` in _ mathematica_. depending on operating systems , the typical directories are + [ cols="^,^",options="header " , ] * * use without installation : one can put _ mathgr _ anywhere and use ` setdirectory ` to specify the load directory . the structure of the _ mathgr _ package.,scaledwidth=90.0% ] the _ mathgr _ package is built in 3 tiers . the structure is illustrated in fig . [ fig : structure ] , where tier 2 depends on tier 1 only , and tier 3 can depend on anything . in this section we introduce the functions provided by each module . to make the notation less cluttered , we have used upper and lower indices in the tensor notations . for example , the inputform of @xmath13 should be input as ] where @xmath14 is a variable . however , once @xmath14 is a variable and assigned to a value like @xmath15 , the indices will have the same replacement and will not work as desired ( unless one does it by purpose ) . thus it is in general recommended to use strings as indices ( or clear and protect the indices before usage if variables are preferred ) . ] @xmath16 , dn["a " ] , dn["b " ] , dn["c"]]!\end{aligned}\ ] ] one can load the typeset.m package ( to be introduced later ) using @xmath17!\end{aligned}\ ] ] after which the inputform is displayed as @xmath13 . one may either copy this @xmath13 to future input to increase readability , or keep the style of inputform . alternatively , one may load most of mathgr packages with one command = ... is the desired output , instead of part of input . ] @xmath18 ! \\ \nonumber & & \verb!out [ ] = mathgr , by yi wang ( 2013 , 2014 ) , https://github.com/tririver/mathgr . ! \\ \nonumber & & \verb!bugs can be reported to https://github.com / tririver / mathgr / issues!\\ \nonumber & & \verb!loaded components tensor , decomp , gr , ibp , typeset , util.!\end{aligned}\ ] ] the module tensor.m provides functions for general tensor calculation and simplification . to make use of tensor.m , we first load the package using @xmath19!\end{aligned}\ ] ] the publicly defined functions and symbols are * _ _ define indices _ _ ] '' is not allowed ( and never necessary in gr ) . ] : this step is optional . as we have encountered previously , we have used up and dn as type identifiers of indices . if this is satisfying , nothing in addition needs to be done . on the other hand , one can define type identifier of indices oneself , for example @xmath20!\end{aligned}\ ] ] here `` \{myu , myd } '' are the new identifiers , `` mydim '' is the dimensions of the manifold on which those indices are defined . `` greekidx '' is a list , which shows the indices to use for dummy indices . other choices are latinidx and latincapitalidx , myu[``a '' ] ] is not allowed . ] . `` red '' is the color in which the indices are typeset . any other color that builtin in mathematica can be chosen . + actually , the identifiers `` \{up , dn } '' are similarly defined in the package : @xmath21!\end{aligned}\ ] ] + as we shall cover in the next sessions , metrics can be defined associated to types of indices . thus defining the type of indices essentially defines manifold or sub - manifold . this provides a mechanism to deal with tensor decomposition from high dimension to lower dimensions . * _ declare tensor symmetry ( declaresym ) _ : this step is also optional . if a tensor has no symmetry , one can use the tensor directly without declaration . on the other hand , for tensors with symmetry , one had better to declare the symmetry such that _ mathgr _ can make use of this symmetry to simplify the tensor . + the symmetry of a tensor @xmath22 can be declared using @xmath23!\end{aligned}\ ] ] here the list of indices are a sequence of up and dn ( or user defined index identifiers ) , representing upper and lower indices respectively . the symmetry could be either symmetric or antisymmetric of some slots , or some generic cyclic symmetries . the grammar for `` symmetry '' here is identical to the one defined in the _ mathematica _ symbolic tensor system thus one can check _ mathematica _ documentation for more details . + for example , @xmath24!\end{aligned}\ ] ] defines a tensor @xmath25 } } t_{abc } ~ , \endaccsupp{}\end{aligned}\ ] ] + where the indices @xmath26 can be any other indices , but upper and lower are distinguished here , as usual in gr . here the tensor is anti - symmetric with permutation of @xmath27 indices . in case of totally symmetric tensor ( declared by symmetric[all ] ) , declaresym also sets the attribute of @xmath22 to be orderless , thus the simplifications are made automatically before sending to the simplification engine . * _ delete tensor symmetry ( deletesym ) _ : the operation declaresym can be undone by deletesym . for example , @xmath28!\end{aligned}\ ] ] removes the symmetry defined above . * _ simplify tensors ( simp ) _ : an expression ( optionally with pre - defined symmetries ) is brought into a unique form by the command simp . this ensures cancellation between different terms in the expression , which can be transformed into each other by using symmetry , or with redefinition of dummy indices . for example , for the tensor @xmath22 defined above , one can run @xmath29 ! \\ \nonumber & & \beginaccsupp{actualtext={t[dn@"a " , dn@"b " , dn@"c"]t[dn@"m " , dn@"n " , dn@"b"]-t[dn@"b " , dn@"a " , dn@"c"]t[dn@"n " , dn@"m " , dn@"b"]}}t_{abc}t_{mnb } - t_{bac}t_{nmb } \endaccsupp { } \verb ! // simp!\\ \nonumber & & \verb!out [ ] = 0 ! \end{aligned}\ ] ] less trivial examples are considered when introducing gr.m , with simplification of curvature tensors . + there is also a list named simphook . this is a list of rules that the user wants to apply before and after simplification . the rules are in the format of rule or ruledelayed , e.g. @xmath30 : > t f[dn@"a " , dn@"b " , dn@"c " ] } } t_{abc } \verb ! :> ! tf_{abc } \endaccsupp { } \verb ! } ! \end{aligned}\ ] ] + note that the simp function does a lot to simplify the expressions , including making assumptions and communicating with the mathematica command tensorreduce . if a tensor expression is expected to contain lots of terms ( say , more than @xmath31 ) after simp , and each term is very simple , simp becomes then overkilling and slow . in this case , one may try to speed up by @xmath32 ! \end{aligned}\ ] ] in this case dummy indices are re - arranged . however , different forms of a tensor are not guaranteed to be brought to the same form . especially , the defined symmetries are not considered ( except for the totally symmetric ones ) . on the other hand , the operation can be order - of - magnitude faster . + note that for mathematica versions lower than 9 , simp is default to fast method because of lack of builtin function tensorreduce . _ partial derivative ( pd ) _ : here we define the partial derivative pd separately , instead of modifying the systemwide partial derivative d. the partial derivative can be called as @xmath33 , dn[index ] ] ! \end{aligned}\ ] ] which displayed as @xmath34 , where @xmath35 is actually \[capitalsampi ] in _ mathematica _ , which looks like @xmath35 ( we try to avoid modifying system symbols such as the real @xmath35 ) . + the derivative pd has linearity and leibniz rules builtin . + internally , the ` pd ` derivative is evaluated into an object named ` pdt ` , with @xmath36 = pd[ ... pd[pd[f , a1 ] , a2 ] , ... ] ! ~,\end{aligned}\ ] ] where ` pdvars ` is a pre - defined orderless function . the reason to convert ` pd ` into ` pdt ` is that , by using ` pdvars ` , the orderlessness of partial derivative is transparent . then it is possible to define general rules on partial derivatives . + by default , the partial derivative acting on all tensors or scalars are nonzero . to define constants with vanishing partial derivative , one has to declare explicitly . for example @xmath37 , _ ] : = 0!\end{aligned}\ ] ] defines both f1 and f2[@xmath38 as constants . + note that this `` define constant '' approach is different from the builtin derivative in _ mathematica _ , where without explicit function dependence , the variable is by default considered as constant . we consider the former to be safer for our purpose , otherwise one may forget to define non - constants . * _ ( anti-)symmetrize tensors ( sym , antisym ) _ : the command @xmath39 ! \end{aligned}\ ] ] symmetrizes the tensor . when indices are not given , all free indices are symmetrized . for example , sym[@xmath40 gives @xmath41 . note that we do not add factors as @xmath42 here . the function antisym does similar things , only that a sign is added in front of each terms , determined by if the permutation is even or odd . * _ the kronecker @xmath43 symbol ( dta ) _ : the simplification of quantities like @xmath44 is automatic . without the need of calling simp , this quantity is directly evaluated into @xmath45 . on the other hand , if there are standalone @xmath46 which can not be simplified , @xmath46 is replaced by @xmath47}_{ab}$ ] . this holdform is released when doing simp . finally , dta@xmath48 gives defaultdim , which denotes dimension of space or spacetime , but without predefined value . note that here we adopt the convention that , the kronecker symbol s indices are not raised or lowered by the metric . for example , @xmath49 is not @xmath50 , instead , @xmath51 if @xmath52 and @xmath53 if @xmath54 , for any metric . this convention is used , as far as i know , widely in modified gravity and cosmic perturbations . also , note that expressions such as @xmath49 and @xmath55 does not raise / lower indices ( e.g. @xmath56 , which has the same outcome as @xmath57 and @xmath58 ) , before ending up this subsection , let us also briefly explain the tensor simplification mechanism under the hood . the @xmath59 function ( unless used with the @xmath60 option ) calls @xmath61 to simplify tensors . and this is why at least version 9 of _ mathematica _ is needed when one needs the full power of tensor simplification . the procedure is described in fig . [ fig : tensor ] . the tensor symmetries defined at @xmath62 are provided as assumptions of @xmath61 . the tensors , on the other hand , are put into coordinate - independent abstract forms before sending into @xmath61 . the procedure to simplify a tensor.,scaledwidth=60.0% ] this package includes definitions which have made use of a metric . to load the package , use @xmath63!\end{aligned}\ ] ] * _ metric definition and usage _ : one may use a metric without any definition . however considering the symmetry and the associated inverse metric , it is convenient to define a metric before usage . moreover , there is also a predefined metric g , which is used by default to calculate metric contractions and curvature tensors . a metric can be defined either globally or locally . the global definition of a metric @xmath64 is ( this metric g predefined in gr.m ) @xmath65!\end{aligned}\ ] ] with this definition , _ mathgr _ declares symmetry for both the metric @xmath66 and the inverse metric @xmath67 . also @xmath68 and @xmath69 are both aliased to @xmath70 . the derivative acting on the inversed metric @xmath71 is replaced using the identity @xmath72 . also , when calculating gr quantities such as ricci scalar r [ ] ( to be introduced below ) , those quantities are calculated with respect to this metric . + when needed , one can define additional metrics , associated with other identifier of index classes ( and thus on a different manifold ) , as @xmath73 ! \nonumber \\ & & \verb!usemetric[myg,{myu , myd}]!\end{aligned}\ ] ] + alternatively , one can use a metric locally , without affecting the global metric . there are two related function usages : + one can define a metric , say @xmath74 , with all metric proprieties but without assigning the system s default metric to @xmath74 , using @xmath75!\end{aligned}\ ] ] + also , one can temporarily define a metric and do some calculation with it , using , e.g. @xmath76!\end{aligned}\ ] ] here the ( some tensors ) are calculated using metric @xmath77 , but outside the withmetric environment , the default metric is still @xmath64 . * _ curvature related symbols and tensors _ : the affine connection @xmath78 ( named affine in _ mathgr _ ) , intrinsic curvature tensors r@xmath79 , r@xmath80 , r@xmath81 , r [ ] , extrinsic curvature tensors k@xmath82 , k [ ] are defined and will be calculated when called . there is also a tensor radm [ ] , which differs from the ricci scalar r [ ] by a total derivative . + for example , @xmath83 // simp ] ! \\ \nonumber & & \verb!out [ ] = ! \frac{3}{4 } g^{ab}g^{cd}g^{ef}\partial_e g_{ac}\partial_fg_{bd } \verb ! + 6 other terms ! \end{aligned}\ ] ] calculates and simplifies the ricci scalar with a temporary metric @xmath64 . note that by default , everything is calculated directly down to metrics . on the other hand , in case one wants to , for example , calculate r [ ] to affine connection and stops , one can locally rename the affine connection to block the calculation going on . for example , @xmath84 // simp ] ! \\ \nonumber & & \verb!out [ ] = ! -g^{ab}\mathrm{gam}^c{}_{ad}\mathrm{gam}^d{}_{cb } \verb ! + 3 other terms ! \end{aligned}\ ] ] + the covariant derivative is also defined , via @xmath85 ! \end{aligned}\ ] ] as is standard in gr . + for example , one can check the second bianchi identity for the curvature tensor @xmath86 , dn@"e " ] + covd[r[dn@"a " , dn@"b " , dn@"d " , dn@"e " ] , dn@"c " ] + covd[r[dn@"a " , dn@"b " , dn@"e " , dn@"c " ] , dn@"d " ] ] // simp } } \verb!withmetric[g , covd[!r_{abcd}\verb ! , dn[e ] ] + covd[!r_{abde}\verb ! , dn[c ] ] + covd[!r_{abec}\verb ! , dn[d ] ] // simp ] ! \endaccsupp { } \\ \nonumber & & \verb!out [ ] = 0 ! \end{aligned}\ ] ] * _ contraction with metric _ : one can definitely use the straightforward way to do contraction . for example , @xmath87 = g[up[#1 ] , up[#2 ] ] f[dn[#1 ] , dn[#2 ] ] & [ unique [ ] , unique [ ] ] ! \\ \nonumber & & \verb!out [ ] = g[up[$6 ] , up[$7 ] ] f[dn[$6 ] , dn[$7 ] ] ! \end{aligned}\ ] ] note that here we have introduced unique variables . ( the `` $ 6 '' and `` $ 7 '' are unique variables as returned by mathematica , where the number could be not 6 , 7 but rather any other numbers . ) otherwise there is a danger that if we use @xmath88 elsewhere with other tensors , the dummy indices of @xmath88 may coincide with the indices of other tensors and cause mistake . to make the unique [ ] function easier to use , we defined uq[n ] to stand for a sequence of unique variables . thus the above equation can be written as @xmath87 = g[up[#1 ] , up[#2 ] ] f[dn[#1 ] , dn[#2 ] ] & @uq[2 ] ! \\ \nonumber & & \verb!out [ ] = g[up[$8 ] , up[$9 ] ] f[dn[$8 ] , dn[$9 ] ] ! \end{aligned}\ ] ] + however , the above expression is still long , and the length will grow quickly with lots of contractions . to save writing , we provide a function metriccontract , which is nothing fundamental , but a shorthand of writing , to allow contraction with the default metric . the above example can be rewritten as @xmath89 = f[dg[1 ] , dg[1 ] ] // metriccontract // simp ! \\ \nonumber & & \verb!out [ ] = ! f_{ab}g^{ab}\end{aligned}\ ] ] where dg and ug are a new type of indices , which is parsed by metriccontract . the same labels are contracted using the default metric . here simp is used to rewrite the dummy indices into familiar ones . + multiple contractions can be calculated similarly , for example @xmath90 = f[dg[1 ] , dg[2 ] ] f[dg[2 ] , dg[3 ] ] f[dg[3 ] , dg[1 ] ] // simp ! \\ \nonumber & & \verb!out [ ] = ! f_{ab } f_{cd } f_{ef } g^{af } g^{bc } g^{de}\end{aligned}\ ] ] the above example calculates @xmath91 , in case only @xmath92 with lower indices are defined . the decomposition module can be loaded by @xmath93!\end{aligned}\ ] ] _ decomposition of tensors into lower dimensional ones ( decomp ) _ : this function converts the dummy indices of @xmath94 dimensional tensors into @xmath95 and @xmath4 dimensional indices respectively . explicit indices can also be used . the explicit indices are marked as ue[n ] and de[n ] for upper and lower indices respectively , where n is a number . the usage is @xmath96 ! \end{aligned}\ ] ] if indices are not given , all dummy indices are decomposed . the rules are of the form , for example , @xmath97 where the lower or upper index is converted into one explicit index de[0 ] or ue[0 ] , and another abstract index . there are a number of predefined decomposition schemes , namely decomp0i[expression , indices ] : convert indices into 0 and i components . the rule is as illustrated above . decomp01i[expression , indices ] : convert indices into 0 , 1 , i components . decomp0123[expression , indices ] : convert indices into all explicit indices , 0 , 1 , 2 , 3 . decomp1i[expression , indices ] : convert indices into 1 , i components . decomp123[expression , indices ] : convert indices into all explicit indices , 1 , 2 , 3 . decompse[expression , indices ] : convert indices into two parts , where both parts has general dimensions . again if the indices are not explicitly specified , all dummy indices are converted . the free indices , if exist , are not touched . in those predefined schemes , the original class identifier of indices are \{utot , dtot}. for example , @xmath98 b[dtot["a " ] ] ] ! \\ \nonumber & & \verb!out [ ] = a[u1["!\alpha\verb!"]]b[d1["!\alpha\verb ! " ] ] + a[u2["a"]]b[d2["a"]]!\end{aligned}\ ] ] here after decomposition , the indices are identified by \{u1 , d1 } , and \{u2 , d2 } ( those index identifiers are defined in decomp.m , with dimensions dim1 and dim2 ) . as another example , @xmath99 b[dtot["a " ] ] ] ! \\ \nonumber & & \verb!out [ ] = a[ue[0]]b[de[0 ] ] + a[up["a"]]b[dn["a"]]!\end{aligned}\ ] ] here the indices are converted into one explicit index and another abstract index . to further ease the calculation , there is a list decomphook , which contains a set of replacement rules to be applied after the decomposition . a frequent use case is to specify the explicit form of the higher dimensional metric using those set of rules , as illustrated in the model file frwadm.m . one may define some `` homogeneous and isotropic background '' quantities with @xmath100:=0 ( * but pd[b , de[0 ] ] is not zero * ) ! \end{aligned}\ ] ] one typical use case for a gr package is to expand an action @xmath101 where if @xmath102 does not have boundary , total derivatives in @xmath103 can be dropped ( e.g. @xmath104 ) . if @xmath102 has a boundary , those total derivatives can be reduced into boundary terms on the boundary @xmath105 . for this purposes , we developed a module to factor @xmath103 into total derivatives and the rest part . as usual , integration - like operations needs more intelligence than derivative - like operations . this module does not guarantee to find the total derivative for complicated cases . nevertheless it works for simple cases and provide convenience for research . in case that @xmath103 is a pure total derivative , the final target result is unique . however , typically , @xmath103 is a total derivative plus some rest part . here depending on use cases , we designed different criteria to try minimizing the rest part : * _ eliminate derivatives on a variable _ : this can be used in a variation principle . for example , @xmath106!\\ & & \verb!ibp[y pd[x , dn["i " ] ] , ibpvar[x ] ] ! \end{aligned}\ ] ] tries to eliminate derivatives acting on @xmath107 , and gives a result @xmath11 = -x pd[y , dn["i " ] ] + pdhold[x y , dn["i"]]!\end{aligned}\ ] ] here pdhold is a function defined to hold total derivatives . one can release this held total derivative by replacing it to pd , or alternatively set it to zero by pdhold[__]=0 if the manifold does not have a boundary . * _ bring the rest part into standard form of a second order action _ : in this case ibp will try to eliminate terms in the rest part with more than two time derivatives on it . and the terms like @xmath108 is transformed into @xmath109pdhold[@xmath110 , de[0 ] ] . the usage is @xmath111 , ibpstd2 ] ! \\ \nonumber & & \verb!out [ ] = -(x^2pd[f , de[0]])/2 + pdhold[(fx^2)/2 , de[0 ] ] ! \end{aligned}\ ] ] * _ leaf count _ : if no criteria is given to ibp , i.e. ibp is called as ibp[expression ] , the leafcount function is applied to compare the rest part . internally , the ibp function works as follows : first , a set of rules with patterns are defined for integration by parts . then every possible rule is applied to the expression and the result is sorted . the one with simplest rest part is chosen and the same set of rules are tried repeatedly on the new result until a fixed point is reached . it is easy to extend the above algorithm such that the rules are applied multiple times before the result is sorted and selected . however in this case the time complexity increase quickly . in typical use cases , ibp may deal with expressions with of order 1000 or more terms ( considering the complexity of the gravitational action ) . in this case multiple - step rules are not realistic . here we have introduced the ibp function motivated by expanding an action . on the other hand , this function can be certainly used for other purposes , as long as total derivatives are wanted . as we mentioned at the beginning of this section , the output of _ mathgr _ can be parsed and brought into a better looking form . for this purpose , the package typeset.m should be loaded . no functions are provided in typeset.m . instead , this module use makeboxes and makeexpression to define the appearance for the tensors . there are also predefined typeset styles for partial derivatives . partial derivative acting on an abstract index is displayed as capital sampi , and time derivative ( w.r.t de[0 ] ) is denoted by dot . to ease some typical calculations , some utilities are provided . those utilities are not directly about tensor calculation but can save some writing for those calculations . the utilities can be loaded using @xmath112!\end{aligned}\ ] ] and provide * _ solveexpr _ : by default , the variable to be inputted to the _ mathematica _ command solve should be atomic or a simple function . expressions with head plus , times and power are not allowed . solveexpr solves this problem . for example , one can use @xmath113 ! \\ \nonumber & & \verb!out [ ] = { { x^2 - > -y } } ! \end{aligned}\ ] ] to find a solution of @xmath114 . to realize this , solveexpr first replace @xmath114 by a unique temporary variable , solve the equation and replace the temporary variable back with @xmath114 . * _ series expansion and coefficients _ : + in _ mathgr _ , the default variable to control orders of perturbations is named eps . in case of a perturbation theory calculation , one multiplies every perturbation variable by eps and expand them together ( an example can be found in frwadm.m ) . the series expansion and extracting the coefficients simply makes use of _ mathematica _ functions series and coefficient . to save some writing , one can use @xmath115!\end{aligned}\ ] ] where n is an explicit integer , to expand expression up to nth order in eps , or @xmath116!\end{aligned}\ ] ] to extract the order eps@xmath117 terms in the expression and disregard all other terms . simplification function simp is called automatically after expansion or extraction of coefficients . to illustrate an explicit use case of the package , here we calculate the cosmic perturbations of inflationary cosmology up to second order . the result is well known for decades @xcite ( for a review with the same notation used here , see @xcite ) . nevertheless to present a standard and familiar calculation for illustration purpose may be more useful for a manual compared with presenting a new and unfamiliar calculation . here we present the input , the explanations and the final results . the intermediate outputs are long and is available in the file resources / mathgr_manual.nb the model specification of a frw universe with adm type perturbations can be loaded by @xmath118!\end{aligned}\ ] ] here the metric @xmath66 is defined by @xmath119 where @xmath120 where @xmath121 . note that for simplicity only the scalar sector is considered in @xmath122 . we have yet one gauge degree of freedom in the scalar sector . we can fix the gauge by either set @xmath123 , or set the inflaton field to be homogeneous and isotropic . we shall consider the latter case as an example . the action up to second order can be calculated with @xmath124:= 0 ( * total derivatives can be neglected here . * ) ! \\ & & \verb!pd[!\phi\verb!\|pd[!\phi\verb ! , de[0 ] ] , _ dn]:= 0 ( * the inflaton perturbation is gauged away . * ) ! \\ & & \verb!s012 = sqrtg ( radm[]/2 + decompg2h[x[!\phi\verb ! ] ] - v[!\phi\verb ! ] ) // ss[2]!\end{aligned}\ ] ] here decompg2h is a function provided in frwadm.m ( may move to more general places in case other models also need this function ) . this function calls decomp0i to decompose a 4-dimensional quantity into 3 + 1 dimensions , where in the 4-dimensional quantities the metric @xmath64 is used and in the 3-dimensional quantities the metric @xmath74 is used . the background equation of motion can be derived by the first order action . for this purpose , we extract the first order action and consider the variation principle : @xmath125 ! \\ & & \verb!solbg = solveexpr[{d[s1 , ! \alpha\verb!]==0 , d[ibp[s1 , ibpvar[!\zeta\verb ! ] ] , ! \zeta\verb!]==0 } , { v[!\phi\verb ! ] , pd[!\phi\verb ! , de@0]^2 } ] ! \\ & & \verb!simphook = union [ simphook , solbg[[1 ] ] ] ! \end{aligned}\ ] ] where on the second line the background equation of motion is solved . note that solveexpr is used because we want to eliminate a composed expression @xmath126 . also note that there are derivatives on @xmath127 in s1 . thus we should first do integration by parts before applying the variation principle d [ ... , @xmath127]=0 . on the third line the background solution is added to simphook . thus it will be automatically applied when simplifying the second order action . here we presented the usage of a new tensor package , _ mathgr _ , which is simple and lightweight , such that people can understand and modify the internal more easily . we shall keep the simplicity of the package . while new functionalities are expected to be added , we shall not add functionalities which significantly increase the complexity of the package , especially for the core parts tensor.m and gr.m . we shall in the future add more comments to the existing code , and add broader coverage for unit tests and integrated tests . those efforts will help for the users who want to hack and fork the package . finally but most importantly , as the package is being tested and used in realistic research , we expect to encounter bugs and provide bug fixes . as always , the result from _ mathgr _ should be checked by independent calculations before being trusted and used in research . we thank j. m. martin - garcia for comments on an earlier version of this paper . this work was supported by the world premier international research center initiative ( wpi initiative ) , mext , japan , a starting grant of the european research council ( erc stg grant 279617 ) , and the stephen hawking advanced fellowship . peter musgrave , denis pollney and kayll lake , http://grtensor.phy.queensu.ca/ ; peter musgrave , denis pollney and kayll lake 1996 `` grtensorii : a package for general relativity '' fields institute communications 15 313 - 318 v. f. mukhanov and g. v. chibisov , `` quantum fluctuation and nonsingular universe . ( in russian ) , '' jetp lett . * 33 * , 532 ( 1981 ) [ pisma zh . fiz . * 33 * , 549 ( 1981 ) ] . y. wang , arxiv:1303.1523 [ hep - th ] .	we introduce the _ mathgr _ package , written in _ mathematica_. the package can manipulate tensor and gr calculations with either abstract or explicit indices , simplify tensors with permutational symmetries , convert tensors from abstract indices to partially or completely explicit indices and convert partial derivatives into total derivatives . frequently used gr tensors and a model of frw universe with adm type perturbations are predefined . the package is built around the philosophy to `` keep it simple '' , and makes use of latest tensor technologies of _
the international linear collider ( ilc ) is a future high energy lepton collider @xcite . it can make precision measurements of the higgs boson , top and electroweak physics . ilc has a capability of exploring new physics as well . to realize these measurements , we should prepare a detector with excellent precision @xcite . in the case of ilc , many events will include jets in their final state . for precise measurements , these jets should be reconstructed as well as possible . we plan to use particle flow algorithm ( pfa ) @xcite to achieve jet energy resolution ( jer ) 3 - 4% . to make full use of pfa , we should separate each shower made from individual particles in jets . thus , we need high granular calorimeter especially in ecal . in this paper , we show basic properties of silicon sensors which should be measured for quality control , and we also report the response to an infrared laser . the silicon sensor chips are produced by hamamatsu ( hpk ) . the sensors are a kind of silicon pin diode , and they have 256 [email protected]@xmath2 pixels in an area of 9@xmath19cm@xmath2 ( figure 1 ) . the thickness of sensors is 320 @xmath4 m . the guard ring can collect surface current of the chips , but it also limit sensitive region of chips . furthermore , from a former test beam result , 1 guard ring structure makes fake signals along it when particles come into near guard ring . in this way , guard ring has advantage and disadvantage , thus we have to decide which design should be used for ild ecal . m ( 1 guard ring ) , or 20 @xmath4 m ( 2 , 4 guard rings).,width=302,height=226 ] there are four types of guard ring structure . currently , we have only 1 guard ring structure in this size . the other structures are provided by smaller chips 3@xmath13 pixels ( 4 guard rings ) , or 4@xmath14 pixels ( no guard ring , 2 guard rings ) . we measured only 1 guard ring type chip , but in the near future we will measure the other chip and compare the properties of each structure . we made two setups for the measurements of basic properties . one of the them is for the measurement of leakage current , the other is for capacitance measurement ( figure 3 ) . we can control temperature and humidity of dark box , thus we are able to assume these two dependence are negligible in both measurements . in our measurements , we set temperature at 27.5 @xmath5c ( or 20.0 @xmath5c for capacitance measurement ) and humidity at 50% . during both measurements , we used copper plate to combine signals from all sensor cells channel at one readout channel ( see figure 4 ) . between sensor and read out circuit are connected by using springs . the operation and data taking were controlled from a pc ( see figure 5 ) . for leakage current measurement , we took data from 0 v to 800 v in 10 v step . each step has 1 second , and if the chip reaches to breakdown voltage we stop the measurement to prevent the chip from getting damaged . as shown in figure 6 , the leakage current is about 100 na with applied bias voltage of 100 - 150 v , and its voltage dependence is very small in this region . we repeated this procedure 10 - 20 times on each chip , and found that the breakdown voltage becomes higher than the previous scan at first 3 - 4 scans . this phenomenon needs to be understood . however , we can conclude the leakage current is quite small and stable at the operation voltage of 100 - 150 v. if a chip has bad channel of high leakage current , we expect that the chip reaches breakdown on much smaller voltage . therefore , this measurement is also important to test the chips before assembly , especially in mass production phase . for capacitance measurement , we applied bias voltage from 0.1 v to 130 v by step of 0.1 v. figure 7 shows good linearity in 5 v to 45 v region . from 0 v to 5 v , the capacitance curve is slightly winding because of its metal - oxide semiconductor structure @xcite . shown in figure 7 , the capacitance saturates at around 60 v , which is consistent of the full depletion voltage of 65 v in the specification . we also estimated that the chip thickness is 318@xmath4 m . this value is also consistent with the specification . to understand the difference between different guard ring structures , we prepared a laser system . we use an infrared laser ( class-3b ) whose wavelength is 1064 nm . photons of this wavelength can pass through the surface of silicon with high transmission to produce a pair of electron - hole in the active area . since infrared photons can not go through electrode , we shoot the laser into the gap ( between a pixel and a pixel , or a pixel and a guard ring ) . at first , we measured the response from a specific pixel to check our system . we used an oscilloscope to see waveform , and we also check the fluctuation of peak value by using the waveform . the waveform taken at 100 v bias voltage is shown in figure 8 . the decay tail was caused by the time constant of the pre - amplifier . we obtained the peak fluctuation value of 3% , shown in figure 9 , which corresponds to the stability of our measurement . we also investigated the response for various bias voltages . the result is shown in figure 10 . in this figure , the response saturates at 80 v , which indicates that the chip reached full depletion . since this saturation is not consistent to the full depletion voltage of 65 v ( obtained from previous measurements ) , we are now investigating the reason of the difference . we established a silicon sensor test system . we can measure basic properties of silicon sensors such as leakage current and capacitance . our first result from these measurements meet the specification . we also investigated the response to an infrared laser . we have a small disagreement for the full depletion voltage between the capacitance measurement and the laser measurement . now we are trying to understand this issue . our final goal is to make a decision on the chip design including guard ring structures , thus we will compare these properties on prototypes of various structures in the near future . 3 t , behnke , james e. brau , b. foster , j. fuster , m. harrison , et al , the international linear collider technical design report - volume 1 : executive summary . 2013 . t , behnke , james e. brau , philip n. burrows , j. fuster , m. peskin , et al . the international linear collider technical design report - volume 4 : detectors . 2013 . m.a . particle flow calorimetry and the pandorapfa algorithm . instrum . and meth . a611 ( 2009 ) 25 - 40 . e.fretwurst , h.herdan , g.lindstroem , u.pein , m.rollwagen , h.schatz , p.thomsen , r.wunstorf , nucl . instr . and meth . a288 ( 1990 ) 1 - 5 .	the international large detector ( ild ) is a proposed detector for the international linear collider ( ilc ) @xcite . it has been designed to achieve an excellent jet energy resolution by using particle flow algorithms ( pfa ) @xcite , which rely on the ability to separate nearby particles within jets . pfa requires calorimeters with high granularity . the ild electromagnetic calorimeter ( ecal ) is a sampling calorimeter with thirty tungsten absorber layers . the total thickness of this ecal is about 24 x@xmath0 , and it has between 10 and 100 million channels to make high granularity . silicon sensors are a candidate technology for the sensitive layers of this ecal . present prototypes of these sensors have 256 [email protected] mm@xmath2 pixels in an area of 9@[email protected] have measured various properties of these prototype sensors : the leakage current , capacitance , and full depletion voltage . we have also examined the response to an infrared laser to understand the sensor s response at its edge and between pixel readout pads , as well the effect of different guard ring designs . + in this paper , we show results from these measurements and discuss future works .
the control of light signal with the magnetic state or conversely the light sensing of the magnetic state is a key issue in the contemporary information technology , as realized in many magneto - optical devices and elements . most of the magneto - optical functions exploit the change of light polarization with the magnetic state , usually generating the light polarization rotation ( kerr or faraday rotation ) . on the other hand , the magnetic - field change of optical absorption itself , i.e. , magnetochromism , is usually very small in spite of the potential demands in spin - electronic science and technology . this is reasonable since the electronic resonance is usually tied to the magnetic state only through spin - orbit interaction in the magnetic medium . some exceptions are found for the materials endowed with inherent strong charge - spin coupling and magnetically controllable competing orders , such as spin - crossover complexes @xcite , colossal - magnetoresistive oxides @xcite , and low - dimensional molecular magnets @xcite . even for these materials , however , a high magnetic field of several tesla is necessary to cause the gigantic magnetochromism . a more promising candidate to host the gigantic magnetochromism is the multiferroic compound , where the electricity and magnetism is inherently coupled @xcite . such a coupling is recently evidenced for a variety of oxide compounds , as typified by the magnetic field control of ferroelectricity in tbmno@xmath9 @xcite and by the electric field control of the magnetization in gdfeo@xmath9 @xcite . as a consequence of such a strong magnetoelectric response , a number of multiferroics may have the potential to show the magnetic resonance in the dielectric constant @xmath10 spectrum , in addition to ( or rather than ) the magnetic permeability spectrum . this collective mode , now termed electromagnon , becomes electrically active and thus can be excited by the light @xmath11 vector . this is contrasted by the case of the conventional magnetic resonance ( the @xmath12 magnon ) driven by the light @xmath13 vector . indeed , recent optical experiments at terahertz frequencies have revealed the emergence of the electromagnon in a family of multiferroics @xcite ; @xmath14mno@xmath9 ( @xmath15 rare earth ions ) @xcite , @xmath14mn@xmath0o@xmath16 @xcite , ba@xmath0mg@xmath0fe@xmath1o@xmath2 @xcite , cufe@xmath17ga@xmath18o@xmath0 @xcite , and ba@xmath0coge@xmath0o@xmath19 @xcite . as the origin of the electromagnons observed in antiferromagnets with non - collinear spin configurations such as @xmath14mno@xmath9 , the exchange - striction mechanism is proposed @xcite : the non - collinear spins ( @xmath20 and @xmath21 on the neighboring @xmath22-th and @xmath23-th sites ) can produce the local polarization @xmath24 in response to @xmath11 component of light , as given by @xmath25 . in fact , the spectral shape of the electromagnon in @xmath14mno@xmath9 can be reproduced by this mechanism @xcite . as only the scalar product @xmath26 on the underlying lattice is relevant , the electromagnon should be insensitive to the direction of the spiral spin plane , as already confirmed by the experiments applying the magnetic field on dymno@xmath9 @xcite . among several magnetoelectrics to host electromagnons , a hexaferrite , ba@xmath0mg@xmath0fe@xmath1o@xmath2 investigated here , occupies a unique position . recent combined studies of terahertz time - domain spectroscopy and inelastic neutron scattering have revealed the emergence of the electromagnon around 0.7 thz ( corresponding to 2.8 mev in photon energy ) @xcite ; the electromagnon becomes active in the conical spin ordered phase when the light @xmath11 vector is set parallel to the propagation vector along [ 001 ] [ see fig . [ figcrstal](d ) ] . at room temperature , ba@xmath0mg@xmath0fe@xmath1o@xmath2 shows the ferrimagnetic order composed of two magnetic sublattice blocks , @xmath27 and @xmath28 [ fig . [ figcrstal](b ) ] with the opposite directions of the in - plane magnetizations . below 195 k , it becomes the proper screw structure , in which the in - plane @xmath27 and @xmath28 spins rotate in proceeding along [ 001 ] [ fig . [ figcrstal](c ) ] @xcite . finally , below 50 k , it forms the longitudinal conical spin structure as the in - plane @xmath27 and @xmath28 spins decline toward [ 001 ] with the conical angle @xmath4 [ fig . [ figcrstal](d ) ] . noticeably , in magnetic fields along [ 001 ] , ba@xmath0mg@xmath0fe@xmath1o@xmath2 yields the large saturation magnetic moment ( @xmath29 @xmath30 per f.u . ; f.u . , formula unit ) . this enables us to easily modify the electromagnon inherent to the conical spin phase [ see the inset of fig . [ fige001pd](a ) ] , as compared to other multiferroics with the antiferromagnetic order ( no net spontaneous magnetization ) . another remarkable characteristic of ba@xmath0mg@xmath0fe@xmath1o@xmath2 is the emergence of the ferroelectricity by controlling the spin cone with magnetic field @xcite : transverse conical structure , as realized by declining the cone axis toward [ 100 ] [ see the inset of fig . [ fige100pd](a ) ] , can produce the ferroelectricity . in such a transverse spiral or conical magnet , the spontaneous polarization stems from the spin - current @xcite or inverse dzyaloshinskii - moriya mechanism @xcite , as expressed by @xmath31 ; here @xmath32 , @xmath33 , @xmath34 are the spin at @xmath22-th site , the unit vector connecting @xmath22- and @xmath23-th sites , and the coupling constant , respectively . this scenario is well proved also for this compound by the detailed analysis of the magnetic structure @xcite . the application of external magnetic field is highly efficient for control of @xmath4 ; @xmath4 can be changed from 90@xmath35 to 0@xmath35 . therefore , the transverse conical structure , where the screw axis is perpendicular to the propagation vector , can host the ferroelectricity along [ 120 ] . according to recent neutron scattering experiments , @xmath4 was found to increase from nearly 0@xmath35 at 50 k to 20@xmath35 at 25 k @xcite . from the viewpoint of the magnetochromism , the magnetically - flexible conical spin structure is expected to produce the dramatic modification of the electromagnon spectrum . for this purpose , we perform the terahertz time - domain spectroscopy and fully map out the spectrum of the electromagnon in a variety of the spin - ordered phases of ba@xmath0mg@xmath0fe@xmath1o@xmath2 by tuning the magnetic field ( up to 7 t ) . the preliminary experiments in the zero and low magnetic field ( up to 0.2 t ) have identified the existence of electromagnon @xcite . here we reveal the gigantic magnetochromism via electromagnons at terahertz frequencies with the fine control of the conical spin order . furthermore , we could determine the magnetoelectric phase diagram of ba@xmath0mg@xmath0fe@xmath1o@xmath2 by conversely using the electromagnon spectra as the probe . the format of this paper is as follows : in sec . ii we describe the experimental setup for terahertz time - domain spectroscopy combined with a split - type superconducting magnet . section iii is devoted to the systematic optical investigations of the electromagnons in a variety of spin - ordered phases tuned by magnetic fields ( from the proper screw to the ferrimagnetic through the conical spin - ordered phases ) , in which we reveal the important role of the conical spin order to produce the electromagnon activity by the measurements in paraelectric longitudinal ( sec . iii a ) and ferroelectric transverse ( sec . iii b ) conical spin - ordered phases . on the basis of the results of sec . iii , we discuss in sec . iv the possible origin of the electromagnon activity in terms of the symmetric - exchange mechanism . the conclusion is given in sec . the single - crystalline samples were prepared by a flux method as described in ref . , and their magnetic and ferroelectric characteristics were all consistent with previous results . the obtained samples were mechanically polished to the thickness of 70 300 @xmath36 m ( thickness of the sample is different in the measurements at respective temperatures so as to gain the signal - to - noise ratio ) . we used the terahertz time - domain spectroscopy in a transmission geometry to directly estimate the complex dielectric constant spectra . our experimental scheme using the znte crystal as a terahertz emitter was reported in ref . . in the present experiments , to further access the low - energy electrodynamics of the electromagnon , especially down to 0.3 thz ( 1.2 mev in photon energy ) , we used the photoswitching sampling technique with a low - temperature - grown gaas ( lt - gaas ) as a terahertz emitter , as schematically shown in fig . [ figthzsetup ] . the femtosecond laser pulses delivered from the mode - locked ti : sapphire laser ( center wavelength of 800 nm ; pulse width of 100 fs ; repetition rate of 82 mhz ) were impinged on the photoswitching device made on the lt - gaas with dipole - antenna . the radiated terahertz pulse was collimated by a pair of off - axis paraboloidal mirrors and focused on the another photoswitching device made on the lt - gaas ; the induced photocurrent was monitored in time - domain by changing the arrival time of the trigger femtosecond laser pulse . in this setup , we inserted the split - type superconducting magnet ( up to 7 t ) in between off - axis paraboloidal mirrors . to eliminate the contribution of the conventional magneto - optical effects , we adopted the voigt geometry , where the light @xmath37 vector is perpendicular to the direction of the magnetic field . the polarization of the radiated terahertz pulse was carefully tuned by a wire - grid polarizer . with use of the fast fourier transformation of the measured terahertz pulse in time domain , we obtained the amplitude and phase spectra with and without the sample . from the obtained complex transmission , we numerically estimated the complex refractive constants @xmath38 , where @xmath39 and @xmath36 are complex dielectric constants and complex magnetic permeability , respectively . as we previously confirmed on the basis of the measurements of the light polarization dependence ( both the electric - field and magnetic - field components of light ) using crystal plates with different crystal orientations@xcite , the contribution of @xmath40 to @xmath41 was negligible in the measured frequency range , namely @xmath42 . in this paper , therefore , we used @xmath43 as a quantity . further details of our estimation procedure were reported in ref . . first , we examine the temperature evolution of the electromagnon in zero magnetic field . the light @xmath11 vector was set parallel to [ 001 ] to induce the electromagnon . in the proper screw phase in the temperature range of 19050 k , there is no signature of electromagnons in the imaginary part of the dielectric constant @xmath44 spectra , as shown in figs . [ fige001spectral](b ) to [ fige001spectral](d ) . below 50 k , where the longitudinal conical spin order evolves [ fig . [ figcrstal](d ) ] , the sharp resonance can be discerned in @xmath45 spectrum and its peak magnitude reaches about 6 , as exemplified by the @xmath45 spectrum at 6 k [ fig . [ fige001spectral](a ) ] . this absorption was assigned to the electromagnon , inherent to the longitudinal conical spin ordered phase , on the basis of the combined measurements of the complete set of the light - polarization dependence as well as of the magnetic excitation spectrum with use of the inelastic neutron scattering @xcite . next , we applied the magnetic field along [ 001 ] at 6 k so as to control @xmath4 of the longitudinal conical structure . in the experiments in magnetic fields @xmath13 , we adopted the voigt geometry to exclude the contribution of the conventional magneto - optical effects such as faraday rotation . owing to the ferrimagnetic nature of ba@xmath0mg@xmath0fe@xmath1o@xmath2 , the magnetization along [ 001 ] increases with magnetic field and yields a large magnetic moment about @xmath29 @xmath30/f.u . at 7 t [ fig . [ fige001pd](a ) ] . the saturated magnitude of @xmath46 is consistent with the anticipated ferrimagnetic order , where @xmath27 and @xmath28 sublattice blocks were perfectly directed along [ 001 ] , as schematically shown in the inset of fig . [ fige001pd](a ) . therefore , we could roughly estimate @xmath4 as a function of magnetic field , as given by @xmath47 [ fig . [ fige001pd](a ) ] . at 6 k in zero magnetic field , @xmath4 was estimated to be 20@xmath35 , which is consistent with the measured value by the recent neutron scattering experiments @xcite . about 3 t , we can induce the ferrimagnetic phase as @xmath4 nearly reaches about 90@xmath35 . in magnetic fields , there is a remarkable change of the electromagnon @xmath45 spectra [ fig . [ fige001spectral](a ) ] ; this is viewed as the gigantic terahertz magnetochromism . with increasing magnetic field , the electromagnon grows in spectral intensity and reaches the maximum around 1 t , as can be seen in fig . [ fige001spectral](a ) . the relative change of the magnitude of @xmath45 at 0.76 thz with and without magnetic field was estimated to be @xmath48 300% in 0.6 t. with further increasing magnetic field , the intensity of the electromagnon decreases and the peak position shifts to the lower energy . the electromagnon resonance disappears above 3 t in the ferrimagnetic ( all spin collinear ) phase . further noticeable magnetochromism can be seen in the @xmath45 spectra at 64 k on the verge of the transition from the proper screw to longitudinal conical spin structures [ fig . [ fige001spectral](b ) ] . the tendency that the intensity of the electromagnon increases with increasing magnetic field is similar to the case at 6 k [ fig . [ fige001spectral](a ) ] , except for the behavior below 0.5 t. in zero magnetic field , the proper screw state is stable , hence no electromagnon is observed . even in weak magnetic fields below 0.5 t , we can induce the electromagnon , via the evolution of the longitudinal conical spin order , as identified in the @xmath45 spectrum around 0.1 t. therefore , near this temperature , the magnetochromism is strongly enhanced ; the relative absorption change yields a gigantic value , @xmath48 500% at 0.6 thz even at 0.6 t. with increasing temperature , the value of the relative change tends to decrease , since the application of magnetic field can hardly induce the longitudinal conical spin phase , as exemplified by the @xmath45 spectra at 90 k [ fig . [ fige001spectral](c ) ] . finally , at 164 k , @xmath45 spectrum is found to be independent of magnetic field ( @xmath49 t ) [ fig . [ fige001spectral](d ) ] . to quantify this trend , we show in fig . [ fige001pd](b ) the contour plot of the spectral weight ( sw ) , as defined by the integration of the optical conductivity ( = @xmath50 , @xmath51 being a permittivity of vacuum ) between 0.4 thz to 1 thz , in the plane of temperature and magnetic field . open circles represent the measured points . the scale color bar is shown at the right side of the figure ; for example , red region indicates the maximum value of the sw . this contour plot clearly reflects the magnetic @xmath52)$ ] phase diagram of ba@xmath0mg@xmath0fe@xmath1o@xmath2 . in the restricted area of the phase diagram , the longitudinal conical spin ordered phase is identified with the enhanced electromagnon spectral intensity dependent on @xmath4 . the present experiments with use of the electromagnon as a probe can provide the full - range phase diagram including the thermal and magnetic - field variations among the proper screw , conical , and collinear structures . in ba@xmath0mg@xmath0fe@xmath1o@xmath2 , we can also induce the transverse conical structure by applying the magnetic field along [ 100 ] . in this phase , the ferroelectricity emerges along [ 120 ] , i.e. , along the direction perpendicular to both the magnetic @xmath53-vector @xmath54)$ ] and the conical axis @xmath55)$ ] , as is evidenced by the measurements of the spontaneous polarization [ fig . [ fige100pd](a)]@xcite . figure [ fige100spectra ] presents the @xmath45 spectra at respective temperatures in magnetic fields @xmath55)$ ] up to 7 t. in zero magnetic field , a signature of the electromagnon at 0.7 thz is discerned only in the longitudinal conical spin phase , e.g. , at 5 k [ fig . [ fige100spectra](a ) ] . when the magnetic field was applied along [ 100 ] at 5 k , the magnetization steeply increases around 0.2 t [ fig . [ fige100pd](a ) ] . accordingly , the spontaneous polarization emerges and reaches the maximum about 80 @xmath36c / m@xmath56 . these are all consistent with the transformation from the longitudinal to transverse conical spin orders , which was directly confirmed by the recent polarized neutron scattering experiments @xcite . the similar magnetochromism is observed in the electromagnon spectra to case of @xmath57 $ ] ( fig . [ fige001spectral ] ) ; @xmath45 increases first with increasing magnetic field and reaches the maximum at around 1 t , and finally , the electromagnon resonance disappears about 3 t , at which the transverse conical spin order transforms to the ferrimagnetic order as evidenced by the reduction of the spontaneous polarization [ fig . [ fige100pd](a ) ] . thus , the emergence of the electromagnon is only limited to the conical spin ( transverse conical for @xmath58 $ ] , and hence ferroelectric in this case ) phase . in turn , we uniquely determine the magnetic phase diagram with use of the electromagnon spectral intensity as a probe . we show in fig . [ fige100pd](b ) the contour plot of the sw ( the integrated optical conductivity from 0.3 thz to 1.2 thz ) in the plane of temperature and magnetic fields . thus , we can identify the region , where sw shows the enhancement in the conical spin arrangement , as colored in red and yellow . the solid lines are the guides to indicate the ferroelectric phase in the phase diagram , as inferred from the recent neutron scattering experiments @xcite , which is consistent with our present data . on the basis of the such mapping of the electromagnon activity in terms of temperature and magnetic field , we confirm again that the conical spin order can act as an effective source of the electromagnons , whose intensity critically depends on @xmath4 . although the detailed theoretical study is needed , the most plausible scenario to explain the above behaviors of the electromagnon is to consider the symmetric exchange - striction mechanism . as extensively discussed in @xmath14mno@xmath9 as the model case @xcite , the non - collinear spins with the symmetric ( heisenberg type ) exchange interaction @xmath59 causes the electromagnon activity through the relationship , @xmath25 , when the local electric dipole is tied to the @xmath60 bond . according to the detailed magnetic structure analysis of ba@xmath0mg@xmath0fe@xmath1o@xmath2 by neutron scattering @xcite , @xmath27 and @xmath28 sublattice blocks are composed of the stacking of fe ions along [ 001 ] , as labeled by 1 " , 4 " , 5 " , 8 " , 11 " , 12 " , 11@xmath61 " , 8@xmath61 " , 5@xmath61 " , 4@xmath61 " , and 1@xmath61 " [ fig . [ figdiscussion](a ) ] . figure [ figdiscussion](b ) shows the reported spin arrangement at respective fe sites within the plane in the ferrimagnetic phase at 294 k @xcite . in this phase , the spins are all collinear . therefore , @xmath62 is perpendicular to @xmath21 and thus @xmath24 in response to @xmath63 $ ] is always zero , as schematically shown in the lower panel of fig . [ figdiscussion](b ) . on the other hand , the spins become non - collinear in the proper screw spin phase below 195 k. the spin arrangement within the plane in this phase is schematically shown in fig . [ figdiscussion](c ) ; the turn angle of @xmath64 at 77 k was taken from ref . . the electromagnon activity would be possible in this phase [ see the lower panel of fig . [ figdiscussion](c ) ] . however , this mechanism can not explain the present experimental results , where there is no remarkable signature of the electromagnon in the proper screw phase [ figs . [ fige001pd](b ) and [ fige100pd](b ) ] . this is partly due to the fact that the in - plane @xmath65 ( scaled with the transition temperature @xmath66 k ) is large compared to the measured photon energy range ( @xmath67 6 mev ) . in this proper screw spin phase of ba@xmath0mg@xmath0fe@xmath1o@xmath2 , @xmath27 and @xmath28 spins lie only within the plane [ fig . [ figdiscussion](c ) ] . on the contrary , spins decline toward [ 001 ] in the longitudinal conical ordered phase below 50 k , where the electromagnon shows up . the spin component projected on to ( 120 ) plane , emphasized at fe 4 " , 5 " , and 8 " sites , is schematically shown in fig . [ figdiscussion](d ) ; the upper , middle , and lower panels stand for the situations , where @xmath4 is @xmath68 , @xmath69 , and 0@xmath35 , respectively . in the case for @xmath7 , there emerges another route to @xmath24 as @xmath62 is not perpendicular to @xmath21 . for example , we can consider @xmath70 produced around fe 5 " site while simply assuming @xmath71 } , 0 , -s^{[001]})$ ] and @xmath72 } , 0 , s^{[001]})$ ] @xcite , the situation for @xmath73 is schematically shown in the middle panel of fig . [ figdiscussion](d ) . we then obtain non - zero component of @xmath70 along [ 001 ] , being proportional to @xmath74}$ ] . in this simple consideration , the intensity of the electromagnon should be scaled with @xmath75 and reach the maximum at @xmath73 [ middle panel of fig . [ figdiscussion](d ) ] . in the longitudinal conical spin phase at 6 k ( fig . [ fige001pd ] ) , for example , the observed sw of the electromagnon reaches the maximum around 1 t , which nearly corresponds to @xmath73 . in the both ferrimagnetic phases induced by magnetic field along [ 001 ] or [ 100 ] , electromagnons become inactive ( figs . [ fige001pd ] and [ fige100pd ] ) as all the spins become collinear within the projection of ( 120 ) plane [ upper panel of fig . [ figdiscussion](d ) ] , being consistent with this mechanism . contrary to the origin of the ferroelectricity , i.e. the antisymmetric exchange interaction described by vector spin chirality @xmath76 , the symmetric exchange interaction @xmath77 acting along [ 001 ] is likely the dominant source of the observed electromagnons . uniquely , such a symmetric exchange interaction depends on @xmath4 . therefore , the magnetic field control of @xmath4 ranging from @xmath78 and @xmath68 can produce the gigantic terahertz magnetochromism via electromagnons as we observed here . in conclusion , we performed the terahertz time - domain spectroscopy for a variety of the magnetically ordered phases of the hexaferrite ba@xmath0mg@xmath0fe@xmath1o@xmath2 , while tuning temperature and magnetic field . the electromagnon appears only in the conical spin ordered phase but irrespective of the longitudinal or transverse conical form , while no electromagnon is observed in the proper screw nor ferrimagnetic phase . the intensity of the electromagnon is found to be highly sensitive to the conical angle tuned by external magnetic fields . such a sensitivity can provide the unique opportunity to determine the magnetoelectric phase diagram in a wide range of temperature and magnetic field . from the technical viewpoint , we clarified that the electromagnon can induce the gigantic magnetochromism at terahertz frequencies ; the estimated relative change of @xmath45 ( @xmath79% ) is remarkably large even in a magnetic field as low as 0.6 t. the important implication of the present work is that the source of the electromagnon is not limited to multiferroics ; electromagnon resonance emerges irrespective of the presence of the ferroelectricity . there are a variety of non - collinear magnets at room temperature , which are candidates to potentially host the gigantic magnetochromism via electromagnons . we thank s. miyahara and t. arima for enlightening discussion . this work was partly supported by grants - in - aid for scientific research ( grant no . 20340086 and 2010458 ) from the mext of japan , and by first program by jsps . j. l. white , c. lee , . gnaydin - sen , l. c. tung , h. m. christen , y. j. wang , m. m. turnbull , c. p. landee , r. d. mcdonald , s. a. crooker , j. singleton , m .- h . whangbo , and j. l. musfeldt , phys . b * 81 , * 052407 ( 2010 ) . @xmath80 and @xmath81 belong to @xmath27 and @xmath28 sublattice blocks , respectively . therefore , to be precise , its magnitude should be different . this also produces the non - zero component along [ 100 ] . even in this case , the present discussion is valid in the context that the non - zero component along [ 001 ] emerges only in the conical spin ordered phase . mg@xmath0fe@xmath1o@xmath2 in zero magnetic field . schematic illustrations of ( a ) hexagonal structure and spin structures in ( b ) ferrimagnetic , ( c ) proper screw , and ( d ) longitudinal conical spin ordered phases . the conical angle @xmath4 is defined in ( d ) . the unit cell can be conveniently divided by two sublattice blocks ; spinel @xmath28 and hexagonal @xmath27 sublattice blocks along [ 001 ] . in the conical spin ordered phase shown in ( d ) , electromagnon becomes active when the light @xmath11 vector is set parallel to [ 001 ] . ] imaginary part of the dielectric constant spectra @xmath44 in magnetic fields up to 7 t , measured at ( a ) 6 k originally in the longitudinal conical spin ordered phase , at ( b ) 64 k near the conical to proper screw transition temperature , at ( c ) 90 k and at ( d ) 164 k originally in the proper screw phase . respective @xmath45 spectra are arbitrarily off - set for clarity ; the off - set revel is indicated by the horizontal solid line . the light @xmath11-vector was set parallel to [ 001 ] and the magnetic field @xmath13 was applied along [ 001 ] in voigt geometry . ] mg@xmath0fe@xmath1o@xmath2 , as determined by the signature of the electromagnon . the magnetic field @xmath13 was applied along [ 001 ] to finely control the conical angle @xmath4 of the longitudinal conical structure , as schematically shown in the inset of ( a ) . ( a ) magnetization along [ 001 ] @xmath82}$ ] and estimate @xmath4 as a function of magnetic field , measured at 6 k. ( b ) contour map of the spectral weight ( sw ) of the electromagnon in terms of temperature and magnetic field . the measured points are indicated by open circles . the magnitude of sw ( see text for the definition ) is shown with the scale color bar . the unit of sw is arbitrary . ] imaginary part of the dielectric constant spectra @xmath44 in magnetic fields up to 7 t measured at ( a ) 5.5 k originally in the conical spin ordered phase , at ( b ) 60.3 k near the conical to proper screw transition temperature , at ( c ) 91 k and at ( d ) 164 k originally in the proper screw phase . respective @xmath45 spectra are arbitrarily off - set for clarity ; the off - set revel is indicated by the horizontal solid line . the light @xmath11-vector was set parallel to [ 001 ] and the magnetic field @xmath13 was applied along [ 100 ] in voigt geometry . ] mg@xmath0fe@xmath1o@xmath2 , as determined by the signature of the electromagnon . the magnetic field @xmath13 was applied along [ 100 ] to induce the transverse conical spin phase , as schematically shown in the inset of ( a ) . in this phase , the ferroelectricity emerges along [ 120 ] , perpendicular to both [ 100 ] and [ 001 ] ( see also fig . [ figcrstal ] ) . ( a ) magnetization along [ 100 ] @xmath83}$ ] and spontaneous polarization along [ 120 ] as a function of magnetic field , measured at 5 k. ( b ) contour map of the spectral weight ( sw ) of the electromagnon in terms of temperature and magnetic field . the measured points are indicated by open circles . the magnitude of sw ( see text for the definition ) is shown with the scale color bar . the unit of sw is arbitrary . the solid line represents the region , where the transverse conical spin structure was identified by the neutron scattering experiments @xcite . in this region , the spontaneous polarization emerges , as shown in ( a ) . ] ( a ) according to the previous neutron scattering experiments @xcite , @xmath27 and @xmath28 sublattice blocks were composed of fe ions , which are labeled by 1 " , 4 " , 5 " , 8 " , 11 " , 12 " , 11@xmath61 " , 8@xmath61 " , 5@xmath61 " , 4@xmath61 " , and 1@xmath61 " . ( b ) in the ferrimagnetic phase at 294 k , the spins at respective sites are collinear with a turn angle of 0@xmath35 ; the reported data were taken from ref . . ( c ) in the proper screw phase at 77 k , the spins are non - collinear with a turn angle of 59.6@xmath35 @xcite . ( d ) possible source of the electromagnons we observed here in the conical spin ordered phase . within the framework of the symmetric - exchange induced striction , electromagnon becomes active when @xmath62 is not perpendicular to @xmath21 . we focus on the fe 4 " , 5 " , and 8 " sites . in the conical phase , the components of the spins projected on to ( 120 ) plane can give rise to the electromagnon activity , as schematically shown in ( d ) . in this case , the electromagnon intensity depends on the conical angle @xmath4 and shows the maximum at @xmath4 of @xmath69 . this can be experimentally realized in magnetic fields . furthermore , there is no chance to produce the electromagnon activity with @xmath4 of @xmath78 or 90@xmath35 . these are all consistent with present experiments . ]	effects of temperature ( 6225 k ) and magnetic field ( 07 t ) on the low - energy ( 1.25 mev ) electrodynamics of the electromagnon , the magnetic resonance driven by the light electric field , have been investigated for a hexaferrite magnet ba@xmath0mg@xmath0fe@xmath1o@xmath2 by using terahertz time - domain spectroscopy . we find the gigantic terahertz magnetochromism via electromagnons ; the magnetochromic change , as defined by the difference of the absorption intensity with and without magnetic field , exceeds @xmath3% even at 0.6 t. the results arise from the fact that the spectral intensity of the electromagnon critically depends on the magnetic structure . with changing the conical spin structures in terms of the conical angle @xmath4 from the proper screw ( @xmath5 ) to the ferrimagnetic ( @xmath6 ) through the conical spin - ordered phases ( @xmath7 ) by external magnetic fields , we identify the maximal magnetochromism around @xmath8 . on the contrary , there is no remarkable signature of the electromagnon in the proper screw and spin - collinear ( ferrimagnetic ) phases , clearly indicating the important role of the conical spin order to produce the magnetically - controllable electromagnons . the possible origin of this electromagnon is argued in terms of the exchange - striction mechanism .
the josephson effect @xcite is one of the key features of superconductors and superfluids . it involves very fundamental properties of these systems and has important applications . the physics of the josephson junctions can be effectively investigated with ultracold gases confined in a double - well potential @xcite . the case of weakly linked bose - einstein condensates ( bec ) has been widely studied . in the seminal papers of smerzi _ @xcite , coupled nonlinear josephson equations for the relative population and the phase difference between the two wells were derived by assuming the system to be described by a superposition of left and right localized condensates ( two - mode model ) governed by the gross - pitaevskii ( gp ) equation ( see also @xcite ) . such nonlinear josephson equations admit solutions in the form of small periodic oscillations , whose period is determined by two key parameters : the mean - field ( on - site ) energy and the tunnelling energy . when the nonlinearity arising from the mean - field interaction exceeds a critical value , the system may exhibit self - trapped solutions with the relative population oscillating around a nonzero value . a large number of theoretical papers have been published along this line and experiments have also been performed @xcite . much less is known about josephson effects in dilute fermi gases . the bogoliubov - de gennes ( bdg ) equations for a two - component superfluid in the crossover from the bardeen - cooper - schrieffer ( bcs ) phase to bec were used in ref . @xcite to describe a stationary supercurrent flowing in the presence of a three - dimensional barrier with a slab geometry ; the current - phase relation and the critical current were studied in the crossover for relatively low barriers , i.e. , height of the barrier smaller than the chemical potential of the superfluid . the same problem was also investigated by means of a density functional approach describing bosonic cooper pairs @xcite ; the equation of state of the gas was included via a suitable parametrization and the order parameter of the superfluid was obtained as the solution of a nonlinear schrdinger equation ( nlse ) . this method gives results in good agreement with the bdg results of ref . @xcite from unitarity to the bec limit . for a double - well potential in the weak link limit ( i.e. , large barriers ) the same density functional can be used to derive coupled nonlinear equations for the relative population and the phase difference analog to those for becs @xcite . a similar nlse has been used to discuss in detail the transition from josephson oscillations to self - trapping @xcite . some open issues are worth considering . first , the applicability of a two - mode model to weakly linked dilute fermi superfluids has been tested so far only within a density functional approach describing a gas of bosonic pairs ( namely cooper pairs , which become molecules in the bec limit ) ; being a generalization of the gp equation , the theory naturally reduces to the two - mode model under the same assumptions as for coupled becs . it is thus interesting to test the two - mode model also within a more microscopic theory like bdg which includes fermionic degrees of freedom . second , the available bdg calculations @xcite and their comparison with the density functional results @xcite are limited to the case of a stationary current through a low and thick barrier , where the flow is almost hydrodynamic and a local density approximation can be applied @xcite ; time dependent simulations with higher and thinner barriers can provide a a more stringent and informative test . finally , the stationary bdg equations does not include bosonic collective modes ( e.g. , phonons ) in the spectrum of excitations and can not address the problem of dynamical instabilities , soliton nucleation , phase slips , etc . , which may occur in a superfluid flow in the presence of a potential barrier . this type of physics can instead be addressed by time - dependent bdg simulations . motivated by the above arguments , we numerically solve the time - dependent bdg equations of a superfluid gas of fermions confined in a box with a square potential barrier at the center . the square barrier is a convenient choice for computational reasons , but the main results of this work would not change by using barriers of different shape . in the limit of weakly linked superfluids we find periodic oscillations whose frequency approaches the prediction of the two - mode model , namely the josephson `` plasma '' frequency @xmath0 @xcite , where the tunneling energy @xmath1 and the on - site mean - field energy @xmath2 are obtained by solving the stationary bdg equations in the same configurations . by increasing the population imbalance we explore the transition from josephson oscillations to self - trapping . we compare the results with those obtained by solving the nlse of refs . @xcite , as well as with the predictions of the coupled nonlinear josephson equations which can be derived from the nlse in the weak link limit . the latter equations turn out to agree surprisingly well with the bdg simulations , provided the parameters @xmath1 and @xmath2 are consistently calculated within the same bdg theory . finally , for large amplitude oscillations and lower barriers the dynamics is complex , involving a combination of josephson - like oscillations , phonons and solitons . we consider a three - dimensional atomic fermi gas at zero temperature with equal populations of two spin components . the interaction between atoms in different spin states is characterized by the s - wave scattering length @xmath3 . in experiments , this parameter can be tuned from small negative ( bcs regime ) to small positive ( bec regime ) values by applying an external magnetic field through a feshbach resonance . on resonance the scattering length diverges and the fermi gas manifests universal properties ( unitarity ) @xcite . we describe the gas in the bcs - bec crossover by means of the bdg equations @xcite @xmath4\left[\begin{array}{c } u_{\eta}({\bf r})\\ v_{\eta}({\bf r } ) \end{array}\right]=\varepsilon_{\eta}\left[\begin{array}{c } u_{\eta}({\bf r})\\ v_{\eta}({\bf r } ) \end{array}\right ] \label{statbdg}\ ] ] where the functions @xmath5 and @xmath6 are the fermionic quasiparticle amplitudes and @xmath7 the corresponding quasiparticle eigenenergy ; @xmath8 is a single - particle grand - canonical hamiltonian for atoms of mass @xmath9 subject to an external potential @xmath10 , and @xmath11 is the chemical potential . the quasiparticle amplitudes obey the normalization condition @xmath12=\delta_{\eta,\eta'}$ ] . the order parameter of the superfluid phase is @xmath13 , where @xmath14 is a coupling constant which accounts for the interaction between atoms ( see eq . ( [ geff ] ) ) ; finally , the atom density is @xmath15 . we also solve the time - dependent version of the bdg equations @xcite : @xmath16\left[\begin{array}{c } u_{\eta}({\bf r},t)\\ v_{\eta}({\bf r},t ) \end{array}\right]=i\hbar\frac{\partial}{\partial t}\left[\begin{array}{c } u_{\eta}({\bf r},t)\\ v_{\eta}({\bf r},t ) \end{array}\right ] \ ; . \label{tdbdg}\ ] ] we use an external potential @xmath17 which depends on the longitudinal coordinate @xmath18 only , consisting of two hard walls placed at @xmath19 and a rectangular barrier of height @xmath20 and width @xmath21 centered at @xmath22 . the lengths @xmath23 and @xmath21 are such that the density and the order parameter of the superfluid are almost constant in the central region of each potential well on the left and right sides of the barrier . in the transverse directions the system is assumed to be uniform ; in practice , we solve the equations in a transverse square box of size @xmath24 imposing periodic boundary conditions for all functions @xmath25 and @xmath26 ; we checked that this value of @xmath27 is large enough to make finite size effects negligible . the average density @xmath28 of the ground state in this potential can be used to define the fermi wave vector of noninteracting fermions of that density , @xmath29 . we use @xmath30 as the unit of length ; we also use the corresponding fermi energy , @xmath31 , as energy unit and @xmath32 as time unit . in the calculations we can either fix the total number of atoms in the box or the density at a point at will . the bdg equations are solved by using the same numerical method as in @xcite . in particular , the solution of the stationary equations ( [ statbdg ] ) are found by a self - consistent iterative procedure . from the solution we can calculate the grand canonical energy @xmath33 of the system as @xmath34 \ ; . \label{eq : energydens}\ ] ] the time - dependent equations ( [ tdbdg ] ) are instead integrated by means of a @xmath35-th order runge - kutta algorithm . the bdg equations require a regularization procedure to cure ultraviolet divergences . here we use the method suggested in ref . @xcite . for this , we choose a cutoff energy @xmath36 sufficiently far above the fermi energy . then , for a given external potential @xmath37 and chemical potential @xmath11 , we define a local fermi wave vector @xmath38 from the relation @xmath39 and a cutoff wave vector @xmath40 from @xmath41 . finally , the regularization of the interaction consists of replacing the bare coupling constant @xmath42 , in the definition of @xmath43 with an effective @xmath14 given by @xmath44 \ ; . \label{geff}\ ] ] the cutoff energy , @xmath45 is chosen large enough to ensure the convergence to cutoff independent results . of a two - component superfluid fermi gas at unitarity and zero temperature obtained by solving the time - dependent bdg equations ( [ tdbdg ] ) . time , in units of @xmath32 , flows from top to bottom . the gas is uniform in the transverse directions and confined between hard walls in the longitudinal direction at @xmath19 with @xmath46 , with a central square barrier of height @xmath47 and width @xmath48 respectively . the number of atoms is @xmath49 . the initial imbalance is produced by adding a constant offset potential @xmath50 at @xmath51 on the left side only ; here we use @xmath52 , which corresponds to an initial relative imbalance @xmath53 . at @xmath54 the offset potential is removed and the system is let to evolve in time . ] let us consider fermions at unitarity ( @xmath55 ) in the presence of a thin ( @xmath56 ) and high ( @xmath57 ) square barrier centered at @xmath22 . we can define the number of atoms on the left , @xmath58 , and right , @xmath59 , as the integrals of the atom density @xmath60 separately in the two regions of negative and positive @xmath18 , respectively . the relative population imbalance can be defined as @xmath61 , where @xmath62 . another key quantity is the phase @xmath63 of the complex order parameter @xmath64 , which can also be different in the two wells . we define the right and left phases as @xmath65 and @xmath66 respectively , and the phase difference as @xmath67 . our simulations start from an imbalanced configuration with @xmath68 . this is obtained by first solving the stationary bdg equations ( [ statbdg ] ) with a small constant offset potential @xmath50 on the left side of the barrier . the ground state solution in such an asymmetric potential is then used as the initial ( @xmath54 ) state in the integration of the time - dependent bdg equations ( [ tdbdg ] ) in the symmetric double - well , after removing @xmath50 . ( @xmath69 , red lines ) and for another simulation with an even smaller initial imbalance ( @xmath70 , black lines ) . ] if the initial imbalance is small ( @xmath71 ) , the time evolution of the density and the order parameter shows clean periodic oscillations . as an example , in fig . [ figjosephson0 ] we show the behavior of the density distribution for an initial imbalance @xmath69 and with a thin and large barrier ( @xmath48 and @xmath47 ) . the evolution of the relative population imbalance @xmath72 and the phase difference @xmath73 is reported in fig . [ figjosephson1 ] . the results for an even smaller imbalance are also shown in the same figure . josephson oscillations between weakly linked superfluids ( @xmath74 ) are characterized by the sinusoidal relation between current and phase difference @xcite : @xmath75 where the quantity @xmath76 has the meaning of critical josephson current . in our case , the current flowing at the barrier position can be easily calculated as @xmath77 . [ curv5d06 ] shows four examples of the current - phase relation obtained in our simulations with different values of the initial imbalance . the upper plots correspond to the simulation of fig . [ figjosephson1 ] . and @xmath78 ) and different initial imbalance , @xmath79 ( a ) , @xmath80 ( b ) , @xmath81 ( c ) , and @xmath82 ( d ) . the red dashed line in panel ( d ) is obtained by solving the nonlinear schrdinger equation ( [ eq : nlse ] ) in the same configuration and for the same initial imbalance . ] if the initial imbalance exceeds a critical value , the system enters into a different dynamical regime , where one of the two wells remains always more populated than the other . the two numbers @xmath58 and @xmath59 oscillate in time , but around unequal mean values . this phenomenon is known as macroscopic quantum self - trapping @xcite . in fig . [ figselftrapping0 ] and [ selftrapping1 ] we show a typical example . the transition from the regime of josephson oscillations and the regime of self - trapping can be visualized by plotting the trajectories in the diagram of the population imbalance _ vs. _ the phase difference . our results for the barrier with @xmath83 and @xmath78 are shown in fig . [ phadiav5d06 ] . josephson oscillations correspond to close trajectories , which become elliptic for small amplitudes , while self - trapping correspond to open trajectories . for the barrier used in these simulations , the transition between the two regime occurs at an initial relative imbalance @xmath84 . but for a larger initial imbalance ( @xmath85 ) , such to cause self - trapping . ] for the same simulation of fig . [ figselftrapping0 ] ( @xmath86 , black line ) and another simulation with an even larger initial imbalance ( @xmath87 , red line ) . ] and @xmath78 ) and different initial imbalance , @xmath70 , @xmath88 , @xmath89 , @xmath90 , @xmath91 , @xmath92 , from the inner ellipse to the outer open trajectory . the red ellipse corresponds to the simulation in fig . [ figjosephson0 ] ; the pink open trajectory corresponds to the simulation in fig . [ figselftrapping0 ] . the transition from josephson oscillations to self - trapping occurs at about @xmath93 . ] the purpose of this section is to show that the above bdg results for small oscillations are well reproduced by josephson junction equations for the two dynamical variables @xmath72 and @xmath73 , provided the barrier is large enough to remain in the weak link regime . in such a situation , the system can be described as composed by two superfluids located in each well and weakly coupled by tunneling ( two - mode model ) . unfortunately , a rigourous derivation of the josephson equations from the bdg equations ( [ tdbdg ] ) within a two - mode approximation is not available . we thus proceed by analogy with the case of bosons where , in the josephson regime , the population imbalance and the phase difference can be seen as canonically conjugates variables entering a classical josephson hamiltonian of the form @xcite @xmath94 the quantity @xmath95 is defined as @xmath96 , where @xmath97 and @xmath98 are the number of bosons on the left and right side of the barrrier , and is assumed to be small . the quantities @xmath2 and @xmath1 have the meaning of on - site energy ( local interaction within each well ) and tunneling energy ( or josephson coupling energy ) , respectively . from ( [ classicaljos ] ) one gets the equations of motion @xmath99 if @xmath100 , the two equations admit harmonic solutions corresponding to josephson oscillations of frequency @xmath101 also known as plasma frequency . these results are valid in the josephson regime where @xmath102 is of order @xmath103 or less , but much larger than @xmath104 ; different regimes are obtained when @xmath105 ( rabi regime ) and @xmath106 ( fock regime ) @xcite . in order to check the applicability of this scheme to the bdg results of the previous section , we need to know how to calculate @xmath2 and @xmath1 within the same theory . we first notice that the tunneling energy @xmath1 can be easily related to the energy difference @xmath107 , where @xmath108 and @xmath109 are the energies of the lowest symmetric and antisymmetric states in the double - well potential with zero imbalance ( @xmath110 ) . in fact , these states have @xmath111 and @xmath112 , respectively , and hence the hamiltonian ( [ classicaljos ] ) gives @xmath113 . moreover we can relate both @xmath1 and @xmath114 to the josephson current @xmath76 . in fact , the number of bosons , i.e. , pairs of fermionic atoms , tunneling through the barrier at @xmath22 per unit time is @xmath115 , so that the current of atoms is @xmath116 , as in eq . ( [ eq : jcurrent ] ) , with @xmath117 . a nice feature of the last relation is that it can be numerically tested by performing two independent calculations . on one hand , the josephson current @xmath76 can be obtained by solving the time dependent bdg equations ( [ tdbdg ] ) : by looking at the current - phase plots , like those in fig . [ curv5d06 ] , the current @xmath76 can be extracted as the maximum of the curve . on the other hand , the energy difference @xmath114 can be calculated by solving the stationary bdg equations ( [ statbdg ] ) for the ground ( symmetric ) state and the lowest antisymmetric state ( see details in the appendix ) . in fig . [ ijdj ] we show the results obtained with @xmath118 particles in a box of size @xmath119 and different barriers . the figure shows that the relation @xmath120 is remarkably well satisfied . extracted from time dependent bdg simulations in the regime of small oscillations and weak tunneling is plotted as a function of the energy difference @xmath107 , between the lowest antisymmetric and symmetric solutions of the stationary bdg equations ( [ statbdg ] ) . all points correspond to @xmath121 , @xmath49 , @xmath119 and @xmath24 , while ( @xmath122 , @xmath123 ) is ( 5,1 ) , ( 5,0.8 ) , ( 5,0.6 ) , ( 6,0.45 ) , ( 5,0.5 ) , ( 4,0.6 ) , ( 4,0.55 ) , for points from bottom - left to top - right . the red line represents the equality @xmath124 . ] the on - site energy @xmath2 accounts for the variation of the interaction energy of the system due to the exchange of particles between the two wells . for a bosonic superfuid in a symmetric well , this parameter is given @xmath125 @xcite , where @xmath126 is the chemical potential and its derivative is calculated at @xmath127 . expressing the same quantity in terms of the chemical potential of the fermionic atoms and the number of atoms , we can write @xmath128 . this quantity can be obtained by solving the stationary bdg equations ( [ statbdg ] ) for different atom numbers in the same double - well . having @xmath1 and @xmath2 , we can finally calculate the plasma period @xmath129 and compare it with the period of the oscillations observed in the time - dependent bdg simulations . the comparison is reported in fig . [ period ] , where we plot @xmath130 ( black solid line ) and @xmath131 ( red dashed line ) as a function of @xmath113 . as one can see , in the limit of small tunneling ( @xmath132 ) , the period observed in the bdg simulations nicely approach the plasma period @xmath131 . these results show that the small oscillations of two weakly coupled fermionic superfluids at unitarity , as obtained with the bdg equations , can be accurately reproduced by a two - mode model for josephson oscillations . , in log - log scale . the black solid line is the period @xmath133 observed in time dependent bdg simulations for @xmath118 atoms in a double - well potential . for each black point the width of the barrier is the same , @xmath78 , while its height decreases from @xmath134 ( leftmost point ) to @xmath135 ( rightmost point ) . the red dashed line is the period @xmath136 of plasma oscillations , where @xmath2 and @xmath1 are calculated by solving the stationary bdg equations ( [ statbdg ] ) , assuming @xmath113 . ] let us now consider larger oscillations and the transition to self - trapping . the classical josephson hamiltonian ( [ classicaljos ] ) does not apply anymore and we may wonder whether nonlinear effects can be properly included in a two - mode model . for bose - einstein condensates governed by the gross - pitaevskii equation , coupled nonlinear josephson junction equations for the number imbalance and the phase were analytically derived by smerzi _ a similar derivation is also available for fermions in the bcs - bec crossover within a phenomenological density functional theory @xcite . this theory is based on the use of the following nonlinear schrdinger equation ( also named density functional gp equation , or extended thomas - fermi equation ) @xmath137\psi({\bf r},t ) \label{eq : nlse}\ ] ] for the order parameter @xmath138 of cooper pairs of mass @xmath139 , with @xmath140 , if @xmath141 is the atom density . the key ingredient of this nonlinear schrdinger equation is the local `` bulk '' chemical potential of cooper pairs , @xmath142 , where @xmath143 is the chemical potential of a uniform fermi gas of density @xmath141 and the second term is the binding energy of the pair . its expression is an input of the theory ; it can be taken from ab initio monte carlo calculations of the equation of state or from the mean - field bdg theory , or different suitable parametrizations . once @xmath144 is given , the nlse ( [ eq : nlse ] ) can be numerically solved for studying stationary and/or time dependent configurations . the advantages and the limits of this approach have been widely discussed in the literature ( see for instance the recent discussion in @xcite , and references therein ) . here we only focus on the fact that , when applied to a double - well potential in the weak link limit , the nlse can be cast into the form of josephson junction equations @xcite . this is done by assuming the order parameter to be a superposition of the left and right parts , @xmath145 having an exponentially small overlap under the central barrier . by inserting this ansatz for @xmath138 into eq . ( [ eq : nlse ] ) , after integration over space and neglecting exponentially small @xmath146 terms , one obtains the equations @xmath147 for the two complex coefficients @xmath148 in region @xmath149 , with @xmath150 . the energy @xmath151 is the sum of @xmath152\psi_i({\bf r})\ ] ] @xmath153 while the coupling term is given by @xmath154\psi_r({\bf r } ) \;.\ ] ] the functions @xmath155 and @xmath156 are real , obey the orthonormality condition @xmath157 and are localized in each of the two wells . in a symmetric system ( i.e. , @xmath158 ) , one has @xmath159 and thus @xmath160 and @xmath161 . by writing @xmath162 and inserting it into eqs . ( [ 2mode1 ] ) and ( [ 2mode2 ] ) , one gets @xcite @xmath163 + \frac{2{\cal k}}{\hbar}\frac{z\cos\phi}{\sqrt{1-z^{2 } } } \label{eq : ge_phase}\end{aligned}\ ] ] where the imbalance @xmath164 and the phase difference @xmath165 are the same already defined at the beginning of section iii . and width @xmath78 , as in fig . [ phadiav5d06 ] . solid lines are the results of time - dependent bdg simulations , while dashed lines are the solutions of the nonlinear josephson equations ( [ eq : ad_dn ] ) and ( [ eq : ad_phase ] ) , with @xmath166 and @xmath167 taken from the solutions of the stationary bdg equations . closed trajectories correspond to josephson oscillations with initial imbalance @xmath69 , while open trajectories correspond to self - trapped states with @xmath86 . ] at unitarity the chemical potential of the uniform fermi gas of density @xmath141 is @xmath168 , where @xmath169 is a universal parameter @xcite . this implies @xmath170 with @xmath171 \int\ ! d{\bf r}\psi_i^{10/3}$ ] , and eq . ( [ eq : ge_phase ] ) becomes @xmath172 + \frac{z\cos\phi}{\sqrt{1-z^{2 } } } \right ) \label{eq : ad_phase}\ ] ] where @xmath173 @xcite . the corresponding classical hamiltonian is @xmath174-\sqrt{1-z^2}\cos\phi \ ; . \label{eq : hclassic}\ ] ] in the limit of small amplitude oscillations ( @xmath100 and @xmath175 ) , the equations of motion ( [ eq : ad_dn ] ) and ( [ eq : ad_phase ] ) reduce to the linear josephson equations ( [ dotk ] ) and ( [ dotphi ] ) provided the two parameters @xmath176 and @xmath177 are related to the on - site interaction energy @xmath2 and the tunneling energy @xmath1 by @xmath178 at this point we are ready to compare our bdg results of section iii with the two - mode model including the nonlinear regime . for each configuration ( i.e. , for each set of parameters @xmath179 ) we can calculate the two energies @xmath2 and @xmath1 by solving the stationary bdg equations as explained in section iv . then we can use them in ( [ eq : kappalambda ] ) to calculate @xmath177 and @xmath176 and solve the nonlinear josephson equations ( [ eq : ad_dn ] ) and ( [ eq : ad_phase ] ) for different values of the initial population imbalance . the results can then be compared with those obtained by solving the time dependent bdg equations ( [ tdbdg ] ) . in fig . [ comv5 ] we show typical results for the imbalance vs. phase diagram , for the same configuration of fig . [ phadiav5d06 ] . the agreement between bdg equations ( solid lines ) and nonlinear josephson equations ( dashed lines ) is remarkably good both in the case of josephson oscillations ( inner ellipse ) and self - trapping ( open trajectories ) . in the bdg simulations the transition between the two regimes occurs at @xmath180 . in the case of the nonlinear josephson equations ( [ eq : ad_dn ] ) and ( [ eq : ad_phase ] ) the same transition is obtained when the energy ( [ eq : hclassic ] ) reaches the critical value @xcite @xmath181 for the parameters of fig . [ comv5 ] , this condition corresponds to @xmath182 , which is again very close to the bdg result . the agreement between bdg equations and nonlinear josephson equations is not restricted to unitarity . we tested that a similar agreement is found also for @xmath183 , both at the bec side ( @xmath184 ) and bcs side ( @xmath185 ) of the bcs - bec crossover . this suggests that the validity of the nonlinear josephson equations ( [ eq : ad_dn ] ) and ( [ eq : ge_phase ] ) is more general than the validity of the nlse ( [ eq : nlse ] ) which is known to be accurate in the bec regime but not in the bcs regime , where it misses the fermionic degrees of freedom . in ref . @xcite it was noticed that , despite this inaccuracy of the nlse , the nonlinear josephson equations can still be used in the whole crossover , provided the tunnelling energy is taken as a phenomenological parameter . our numerical results show that the same nonlinear josephson equations are a very good approximation of the weak link limit of the bdg equations , the parameters @xmath2 and @xmath1 being consistently calculated within the same bdg theory . the difference between nlse and bdg equations can be appreciated by looking at fig . [ nlsevsbdgejij ] , where we plot the results for the maximum josephson current , @xmath76 , together with the energy difference @xmath114 . the quantity @xmath76 is extracted from time dependent simulations , either solving the bdg equations ( [ tdbdg ] ) ( red solid line ) or the nlse ( [ eq : nlse ] ) ( upper dashed line ) , while @xmath114 is calculated from the corresponding stationary ( time independent ) equations ; in eq . ( [ eq : nlse ] ) we use the mean - field equation of state ( mf eos ) for the local chemical potential @xcite . as discussed in section iv , in the weak link limit , where the nonlinear josephson equations are expected to hold , the quantity @xmath114 should be equal to twice the tunneling energy @xmath1 and one should find @xmath186 . this is clearly the case for bdg equations where @xmath114 ( black solid line ) and @xmath187 ( red solid line ) are almost indistinguishable in the whole crossover , the small difference in the bec limit being likely due to the finite cutoff energy in the bdg calculations , which becomes a more critical parameter as @xmath188 increases . conversely in the case of nlse , the two quantity are significantly different and the critical current @xmath76 is increasingly larger than the bdg prediction in the bcs limit . the difference can be seen also in fig . [ curv5d06 ] where we show an example of josephson oscillations at unitarity as obtained by solving eq . ( [ eq : nlse ] ) ( dashed line ) and eq . ( [ tdbdg ] ) ( solid line ) for the same configuration . the fact that @xmath76 is larger in the nlse than in bdg equations is well known and is simply due to pair - breaking processes which are included in bdg @xcite but are absent in the nlse . this effect was already discussed in ref . @xcite in a regime of wider ( @xmath189 ) and lower ( @xmath190 ) barriers . here , on purpose , we have chosen thinner barriers , i.e , @xmath21 of the order or less than @xmath30 , in order to test the applicability of the two - mode model to cases where density and phase variations occur on the lengthscale of the inverse fermi wave vector , such that the local density approximation becomes questionable and fermionic degrees of freedom might play a role . our results indicate that , at least in the weak link limit and within a mean - field theory , the dynamics is still dominated by tunneling of bosonic pairs and is suprisingly well described by the nonlinear josephson equations ( [ eq : ad_dn])-([eq : ge_phase ] ) . and maximum josephson current @xmath76 calculated with the bdg equations ( solid lines ) and the nlse ( dashed lines ) , as a function of the interaction strength @xmath188 . the parameters of the barrier are @xmath48 and @xmath47 , and the number of atoms @xmath49 . ] the situation is rather different when the coupling between the two wells is strong . an example is shown in fig . [ v1d06den ] , where we plot the density @xmath191 in a bdg simulation with a low barrier ( @xmath192 and @xmath78 ) and large initial imbalance . the josephson current through the barrier is strongly coupled to the collective motion of the gas in the two wells . one can distinguish a density wave bouncing back and forth with a velocity of the order of the sound speed in a unitary fermi gas with the same average density , @xmath193@xcite . in addition , at about @xmath194 , when the density under the barrier almost vanishes , a grey soliton is nucleated . the soliton appears as a density depletion travelling leftward ( dashed line ) at a velocity smaller than the speed of sound . the phase of the order parameter has a variation of the order of @xmath195 across the soliton . in the case of an infinite system , this mechanism of soliton nucleation induces a dissipation of the superfluid current due to phase slip @xcite . in our confined double - well system , solitons and collective sound - like waves are coupled by nonlinear mixing and eventually lead to a decay of the initial josephson oscillation . at unitarity , obtained by solving the time - dependent bdg equations ( [ tdbdg ] ) , as in fig . [ figjosephson0 ] , but for a lower barrier ( @xmath192 and @xmath78 ) . the initial imbalance is @xmath196 . solid and dashed lines represent the propagation of a sound - like density wave packet and a grey soliton , respectively . ] two weakly linked fermi superfluids in a double well geometry exhibit dynamical regimes of josephson oscillations and self - trapping . the nonlinear josephson equations that describe both regimes are the analog of those for bose superfluids described by the gp equation @xcite ; indeed the fermionic versions of such equations were previously derived starting from a generalized gp equation @xcite . here we show that the same nonlinear josephson equations are in remarkable agreement with time - dependent bdg simulations , with the on - site energy , @xmath2 , and the tunneling energy , @xmath1 , consistently calculated within the same bdg theory . such an agreement could not be foretold _ a priori _ since , on one hand , a formal derivation of the nonlinear josephson equations from the bdg equations is yet missing and , on the other hand , the role played by fermionic degrees of freedom in the dynamics of a superfluid subject to density and phase variations on the lengthscale of the inverse fermi wave vector is largely unkown . our results indicate that the dynamics at the weak link is dominated by tunneling of bosonic pairs , with the caveat that the critical josephson current @xmath76 in the bcs side of the crossover is determined by pair - beaking processes . for lower barriers and large tunneling , the dynamics involves the collective motion of the superfluid in the two wells and possible phase slip processes due to the nucleation of solitons . our predictions complement those of ref . @xcite and are intended to stimulate new experimental investigations with ultracold fermi gases in double - well potentials . we thank s. stringari , l.p . pitaevskii , a. recati for fruitful discussions . p.z is particularly indebted to r.g.scott , whose computational expertise was of great help at the beginning of this work . support of erc , through the qgbe grant , and of provincia autonoma di trento is acknowledged . in this appendix we provide some details about the calculations of the energy difference , @xmath197 , where @xmath108 is the energy of the ( symmetric ) ground state of the gas and @xmath109 is the energy of the lowest anti - symmetric state with the same number of particles @xmath198 in the box . here symmetric and anti - symmetric refer to spatial reflection in the @xmath18-direction around @xmath199 . the anti - symmetric state corresponds to the solutions of eq . ( [ statbdg ] ) exhibiting a @xmath195 phase jump in the order parameter @xmath43 when crossing the center of the box . for weakly coupled superfluids the quantity @xmath114 is directly related to the josephson tunnelling energy , @xmath1 as already seen in section iv . , divided by the area @xmath200 , as a function of the barrier height @xmath20 for different widths @xmath21 . right : the same quantity as a function of the barrier width @xmath21 for different values of height @xmath20 . all the other parameters are the same as in the previous figure . ] since the quantity @xmath114 is typically much smaller than the energies @xmath109 and @xmath108 , we must take care of all possible sources of numerical inaccuracy , in particular those introduced by the finite cutoff energy and the finite box . we have first checked that , at unitarity , without barrier and in the limit of large @xmath23 , the energy @xmath108 converges to the energy of a uniform infinite gas : @xmath205 , where @xmath206 is the value of the bertsch parameter in bdg theory @xcite . in the same situation , the quantity @xmath114 measures the cost in energy associated to creation of the density depletion and the nodal structure of the order parameter at the box center , corresponding to a dark soliton @xcite . the energy per unit surface of a planar dark soliton at unitarity is @xmath207 and bdg theory gives @xmath208 @xcite . we have checked that both analytic values are reproduced within an accuracy of about @xmath209% , which is enough for our purposes . our results for @xmath114 as a function of @xmath20 and @xmath21 are shown in figs . [ figejvd - d2v2 ] and [ figejvd ] . all results in these figures are obtained by using @xmath201 , @xmath202 , @xmath203 , and @xmath204 . as expected , @xmath114 approaches the same value in the limit of vanishingly small barrier ( i.e , for @xmath210 at finite @xmath21 ) . this value is @xmath211 , which is slightly larger than the energy of dark soliton in an infinite system , due to the finite box sixe . for @xmath21 of the order of @xmath30 , the quantity @xmath114 is rapidly decreasing when @xmath20 increases . the case of large barriers and small tunnelling ( weak link ) is where the physics of the josephson effect is expected to manifest . finally we note that for @xmath20 much smaller than @xmath212 the quantity @xmath114 tends to be a constant value when @xmath213 . a simple explanation of this behavior is obtained by considering that , for a very wide and low barrier , the effect of the barrier is that of lowering the `` bulk '' density in the central region of the box . if @xmath21 is larger that the soliton width , which is of the order of a few @xmath30 , this effect can be accounted for by calculating the energy of a dark soliton in uniform gas of reduced density . further increasing the width of the barrier has no effects on the soliton energy and hence @xmath114 remains constant . th.anker , m.albiez , r.gati , s.hunsmann , b.eiermann , a.trombettoni , and m.k.oberthaler , phys . * 94 * , 020403 ( 2005 ) ; m.albiez , r.gati , j.foelling , s.hunsmann , m.cristiani , m.k.oberthaler , phys . lett . * 95 * , 010402 ( 2005 ) ; t.zibold , e.nicklas , ch.gross , and m.k.oberthaler , phys . lett . * 105 * , 204101 ( 2010 ) . de gennes , _ superconductivity of metals and alloys _ , ( benjamin , new york , 1966 ) ; a.j . leggett , in _ modern trends in the theory of condensed matter _ , edited by a. pekalski and r. przystawa ( springer - verlag , berlin , 1980 ) ; m. randeria , in _ bose einstein condensation _ , edited by a. griffin , d. snoke , and s. stringari ( cambridge , cambridge , england,1995 ) . here the notation is the same as in ref.@xcite except for the meaning of @xmath198 which , in the present work , is the number of fermionic atoms , while in @xcite is the number of bosonic dimers . in the context of nlse the two numbers differ only by a factor two .	we investigate the behaviour of a two - component fermi superfluid in a double - well potential . we numerically solve the time dependent bogoliubov - de gennes equations and characterize the regimes of josephson oscillations and self - trapping for different potential barriers and initial conditions . in the weak link limit the results agree with a two - mode model where the relative population and the phase difference between the two wells obey coupled nonlinear josephson equations . a more complex dynamics is predicted for large amplitude oscillations and large tunneling .
transient black hole binaries ( bhbs ) display outbursts exhibiting several states , characterized by both spectral and timing properties ( e.g. belloni et al . 2005 ; remillard & mcclintock 2006 ; belloni 2010 ; gilfanov 2010 ) . during the outburst , sources typically follow a q shaped , anti - clockwise track on a plot of x - ray flux versus spectral hardness ratio ( hardness - intensity diagram : hid ) , with the quiescent state occupying the bottom right corner . the initial transition from hard ( lhs ) to soft ( hss ) , via intermediate states , occurs when the power low component of the spectrum is observed to soften ( photon index @xmath2 1.72.4 ) and a disc blackbody component ( peaking in soft x - rays ) becomes increasingly prominent . a power spectral analysis of the rapid variability reveals a quasi - periodic oscillation ( qpo ) , which shows up as narrow harmonically related peaks , superimposed on broad band continuum noise . the qpo fundamental frequency is observed to increase from @xmath3 0.110 hz during the transition from the hard state , after which the x - ray emission becomes very stable in the soft state . power spectral evolution correlates tightly with spectral evolution , with all the characteristic frequencies increasing with spectral hardness ( e.g. wijnands & van der klis 1998 ; psaltis , belloni & van der klis 1999 ; homan et al . 2001 ) . qpos observed coincident with broad band noise are defined as type - c qpos ( remillard et al . 2002 ; casella et al . type - b qpos ( wijnands , homan & van der klis 1999 ) , typically with a frequency of @xmath3 610 hz , are observed in the intermediate state when the broad band noise suddenly disappears . these features quickly evolve into type - a qpos ( wijnands , homan & van der klis 1999 ) , which are broader and weaker . since the sudden suppression of the broad band noise hints a large physical change in the system , intermediate state observations displaying type - c qpos are classified as hard intermediate state ( hims ) and those displaying type - a or b qpos as soft intermediate state ( sims ) . additionally , a large radio flare , indicative of a jet ejection event , is often observed to be coincident with the onset of the sims ( fender , belloni & gallo 2004 ) , although this is not always exact ( fender , belloni , & gallo 2005 ) . + the spectral and timing properties of bhbs can be described by _ the truncated disc model _ ( e.g. esin , mcclintock & narayan 1997 ; done , gierliski & kubota 2007 ) where an optically thick , geometrically thin accretion disc which produces the multi - temperature blackbody spectral component ( shakura & sunyaev 1973 ) truncates at some radius , @xmath4 , larger than the innermost stable circular orbit ( isco ) . in the region between this truncation radius @xmath4 and an inner radius @xmath5 ( @xmath6 ) , accretion takes place via a hot , optically thin , geometrically thick accretion flow ( hereafter inner flow ) . compton up - scattering of cool disc photons by hot electrons in the flow produces the power low spectral component ( thorne & price 1975 ; sunyaev & truemper 1979 ) . in the hard state @xmath4 is large ( @xmath7 , where @xmath8 is a gravitational radius ) , so only a small fraction of the disc photons illuminates the flow , giving rise to a weak direct disc component and hard power law emission . as the average mass accretion rate increases during the outburst , @xmath4 decreases , so more direct disc emission is seen and a greater luminosity of disc photons cool the flow , resulting in softer power law emission . when @xmath4 reaches the isco , the direct disc emission completely dominates the spectrum and the transition to the soft state is complete . + this scenario is the framework of the propagating fluctuations model propfluc ( ingram & done 2011 , 2012 , hereafter id11 , id12 ; ingram & van der klis 2013 , ik13 ) , a model that can reproduce power density spectra by combining the effects of the propagation of mass accretion rate fluctuations in the inner flow ( lyubarskii 1997 ; arevalo & uttley 2006 ) , responsible for generating the broad band noise , with solid - body lense - thirring ( lt ) precession of this flow ( fragile et al . 2007 ; ingram , done & fragile 2009 ) , producing qpos . mass accretion rate fluctuations are generated throughout the inner flow , with the contribution to the rms variability from each region peaking at the local viscous frequency ( e.g. lyubarskii 1997 ; churazov , gilfanov & revnivtsev 2001 ; arevalo & uttley 2006 ) , thus the fast variability originates from the inner regions and the slow variability from the outer regions . as material is accreted , fluctuations propagate inwards , modulating the faster variability generated in the inner regions . emission is thus highly correlated from all regions of the flow , giving rise to the observed linear rms - flux relation ( uttley & mchardy 2001 ; uttley , vaughan & mchardy 2005 ) . + in this paper we present a spectral and timing analysis of the source maxi j1543 - 564 during its 2011 outburst . the source , discovered by maxi / gsc ( the gas slit camera of the monitor of all - sky x - ray image ; matsuoka et al . 2009 ) on may 08 2011 ( negoro et al . 2011 ) , was first analyzed by stiele et al . their analysis showed that the outburst evolution follows the usual bhbs behavior , the exponential flux decay is interrupted by several flares , and during the transition from lhs and hss a type - c qpo is observed . looking at other wavelengths , miller - jones et al . ( 2011 ) report the detection of radio emission at mjd 55695.73 . in this work , we analyze the spectral and timing properties of the source in different energy bands and we use the power density spectra of the rising phase of the outburst to systematically explore for the first time the capabilities of propfluc . + we analyzed data from the rxte proportional counter array ( pca ; jahoda et al . 1996 ) using 99 pointed observations collected between 10 may and 30 september 2011 . each observation consisted of between 3004750 s of useful data . + we used standard 2 mode data ( 16 s time resolution ) to calculate a hard color ( hc ) as the 16.020.0 / 2.06.0 kev count rate ratio and define the intensity as the count rate in the 2.020.0 kev band . all the observations were background subtracted and all count rates were normalized by the corresponding crab values closest in time to the observations . + we used the @xmath3 125 @xmath9s time resolution event mode and the @xmath3 1 @xmath9s time resolution good - xenon mode data for fourier timing analysis . we constructed leahy - normalized power spectra using 128 s data segments and 1/8192 s time bins to obtain a frequency resolution of 1/128 hz and a nyquist frequency of 4096 hz . after averaging these power spectra per observation , we subtracted the poisson noise using the method developed by klein - wolt et al . ( 2004 ) , based on the expression of zhang et al . ( 1995 ) , and renormalized the spectra to power density @xmath10 in units of @xmath11 / hz . in this normalization the fractional rms of a variability component is directly proportional to the square root of its integrated power density : @xmath12 % . no background or dead - time corrections were made in computing the power spectra . this procedure was performed in 4 different energy bands : 2.874.90 kev ( band 1 ) , 4.909.81 kev ( band 2 ) , 9.8120.20 kev ( band 3 ) , and the full 2.87-20.20 kev ( band 0 ) . the power spectra were fitted using a multi - lorentzian function in which each lorentzian contributing to the fit function is specified by a characteristic frequency @xmath13 ( belloni , psaltis & van der klis 2002 ) and a quality factor @xmath14 , where fwhm is the full width at half maximum and @xmath15 is the centroid frequency of the lorentzian . all the power spectra shown in this paper were plotted using the power times frequency representation ( @xmath16 ) , in order to visualize @xmath17 as the frequency where the lorentzian s maximum occurs . the light curve of the source is shown in fig . [ fig : tid]@xmath18 , where the 2 - 20 kev intensity is plotted versus time ( mjd ) for each pointed observation . + we subdivided the evolution of the outburst in 5 parts . in the first part of the outburst ( mjd 5569155696 ) the source rises to maximum intensity ( @xmath3 68 mcrab ) in 5 days from the beginning of the rxte observations . the second part ( first grey area , mjd 5569655713 ) is characterized by an intensity decay that is not smooth , but interrupted by 4 additional peaks with intensities between @xmath3 47 and @xmath3 58 mcrab . the third part ( mjd 5571355725 , between the two grey areas ) does not show any intensity peak but only a gradual decay . the following period ( mjd 5572555744 , second grey area ) is characterized by a broad maximum and several additional intensity peaks ( between @xmath3 34 and @xmath3 42 mcrab ) less luminous compared to those of the first grey area . finally , the last part ( mjd 5574455834 ) consists of a relatively smooth decay until the end of the observations . + , b ) rms [ % ] , c ) and hard color [ crab ] versus time for the 99 pointed observations . the grey rectangular areas indicate 5 time intervals characterized by different long - term luminosity variability . data points are plotted with 1@xmath19 error bars . ] [ fig : hid ] shows the hardness-intensity diagram ( hid ) , where the average intensity of each observation is plotted versus the hc . the source follows a counterclockwise path , starting and ending in the right ( hard ) part of the diagram at different luminosities . this is the usual behavior observed for black hole outbursts . + in order to better follow the spectral evolution of the source along the outburst , we also plotted in fig . [ fig : tid]@xmath20 the hc versus time . + in the first observation the source is harder than crab ( hc = 1.71 ) and in the following 6 observations softens continuously , while at the same time its intensity increases from @xmath3 24 to @xmath3 68 mcrab . for the remaining observations the source remains in the soft part of the hid ( [email protected] ) except for the very last observation , where it goes back to a color harder than crab ( hc = [email protected] ) . + as can be noted in fig . [ fig : tid]@xmath20 , the transitions between hard and soft spectrum happen on short time scales ( @xmath3 10 days ) compared to the time spent by the source in the soft state ( @xmath3 125 days ) . however , while the initial transition from hard to soft state is simultaneous with a quick change in intensity ( + 188@xmath23 ) , the final transition ( last observation ) from soft to hard spectrum is characterized by a fractional intensity change of only + 16@xmath23 , i.e. increasing when the source gets harder . + error bars are plotted for the hard color . ] the 1/12910 hz rms values as computed from the power spectra in band 0 are reported in fig . [ fig : tid]@xmath24 . the first 5 observations , during which the source rapidly becomes softer and brighter are characterized by rms values of @xmath3 1927@xmath23 . in the remaining observations the rms values are between @xmath3 2@xmath23 and @xmath3 10@xmath23 with few exceptions . + integrated rms is systematically higher for higher energies . from the beginning of the observations , as the intensity increases , integrated rms decreases independently from photon energy , but the rms decrease trend is different between energy bands . in order to better show these differences , we plotted in fig . [ fig : rmsall ] the total fractional rms of the first 7 observations for all the energy bands . band 1 ( red ) shows a smooth and continuous rms decrease with time , while in band 2 ( green ) and band 3 ( blue ) , the rms decrease is characterized by a jump between observations # 5 and # 6 ( @xmath25rms @xmath3 9@xmath23 in band 2 , @xmath25rms @xmath3 11@xmath23 in band 3 ) . observation # 6 is also characterized by the detection of radio emission , indicated by the orange arrow . + , its harmonic @xmath26 , a broad band noise component @xmath27 , and another broad component at lower frequency @xmath28 . ] 7 showing a type - b qpo . ] only in the first 7 observations we detect qpos ( @xmath29 ) and/or broad band components ( @xmath30 ) in at least some energy bands . we used the power spectrum of the fifth observation in band 0 ( mjd 55694.884 , fig . [ fig : psex ] ) as a reference to identify four significant ( @xmath31 , single - trial ) components : a main qpo @xmath32 , its harmonic @xmath26 , a broad band noise component @xmath27 , and another broad band component @xmath28 at lower frequency . in our analysis we reported all components with single trial significance @xmath33 and additionally those components with significance between 2@xmath19 and 3@xmath19 that could be identified as @xmath32 , @xmath26 , @xmath27 , or @xmath28 . table [ tab : fmax ] shows @xmath34 , q , rms , significance ( @xmath19 ) and reduced @xmath35 for every fitted component in the 7 observations analyzed ( # 17 ) for all the energy bands . we also report the 99.87% upper limits calculated fixing @xmath34 and q to values equal to the most significant corresponding component between the energy bands fitted in the same observation . empty lines mean that no components were fitted and no upper limit could be determined . [ fig : fmax]@xmath36@xmath37 show the frequencies of the fitted qpos ( triangles for @xmath32 , diamonds for @xmath26 ) , broad band components ( squares for @xmath27 , circles for @xmath28 ) , and unidentified narrow ( @xmath29 ) components ( pentagons ) , and their rms versus time in band 0 , respectively . solid symbols indicate significant components and open symbols components with significance between 2@xmath19 and 3@xmath19 . the 23@xmath19 unidentified component of observation # 6 ( see table [ tab : fmax ] , bottom ) is included in our plot because its characteristic frequency matches with the subharmonic frequency of the identified component @xmath32 . similarly , two 23@xmath19 unidentified components fitted in observation # 7 ( fig . [ fig : typeb ] ) were reported , as one matches with the subharmonic frequency of @xmath32 , and the other with @xmath38 . squares and circles were slightly shifted to the right for clarity . + always referring to band 0 , in the first 5 observations one significant low frequency qpo ( @xmath32 ) was fitted for each spectrum and only the third observation shows a significant harmonic ( @xmath26 ) . the @xmath32 frequency increases with time from @xmath3 1.1 hz to @xmath3 5.8 hz while its rms decreases from @xmath3 17@xmath23 to @xmath3 10@xmath23 ( see table [ tab : fmax ] ) . observation # 7 shows a significant qpo with @xmath17 = 4.7 hz ( fig . [ fig : typeb ] ) . the peak characteristics ( @xmath17 = 4.7 hz , @xmath39 = 9 , rms @xmath3 4.8 ) of and the low 1/12810 hz rms ( @xmath3 7.2% ) associated with this qpo , are characteristics of type - b qpos ( e.g. casella et al . considering also the 23@xmath19 qpo fitted in observation # 6 ( @xmath17 = 5.7 hz , @xmath19 @xmath3 2.5 ) , in observations # 67 @xmath32 frequency and rms are not anti - correlated anymore . the characteristic frequency of @xmath32 decreases from @xmath3 5.8 hz to @xmath3 4.7 hz while the rms still decreases from @xmath3 6@xmath23 to @xmath3 5@xmath23 . + one significant broad band component ( @xmath27 ) with @xmath17 in the interval @xmath3 24 hz was fitted in observations # 16 . the rms of this component decreases with time ( from 20@xmath23 to 9@xmath23 ) , with a clear decreasing trend observable only in observations # 56 , while its @xmath17 remains almost in the same frequency range ( around 3 hz ) . in observations 57 we fitted another broad band component ( @xmath28 ) characterized by an increasing @xmath17 ( from observation # 5 to # 7 ) in the interval @xmath3 0.060.66 hz and rms between 2% and 4% . + the timing features in the other energy bands are reported in fig . [ fig : fmax]@xmath18@xmath40 . similarly to panels @xmath36@xmath37 , plots @xmath18@xmath24 , @xmath20@xmath41 , and @xmath42@xmath40 show frequency and rms evolution for power spectral components fitted in bands 1 , 2 , and 3 respectively . no significant characteristic frequency shift was detected between energy bands in any power spectral component , while the rms values are systematically higher for higher energies ( table [ tab : fmax ] ) . in band 1 @xmath32 frequency increases with time ( from @xmath3 1.1 hz to @xmath3 6.5 hz ) in the first 6 observations while no - significant qpo was fitted in observation # 7 . the behavior of @xmath32 characteristic frequency in bands 2 - 3 is mostly identical to band 1 for observations # 15 , but we observe some differences in observations # 67 . the 23@xmath19 qpo ( @xmath19 @xmath3 2.2 ) fitted in observation # 6 ( band 2 ) seems to break the anti - correlation between frequency and rms shown in observations # 15 , but in band 3 the qpo frequency error bar is too big to infer any trend . however , the anti - correlation is evident in observation # 7 , where a significant qpo was fitted in bands 2 - 3 with lower characteristic frequency compared to observation # 5 . the rms of @xmath32 in band 1 decreases as the qpo frequency increases , but in bands 23 this trend is progressively weaker . indeed , in band 3 the @xmath32 rms slightly oscillates around @xmath3 17@xmath23 and @xmath3 11@xmath23 in the first 5 observations and decreases to @xmath3 11@xmath23 only in the last 2 observations . + the broad band component ( @xmath27 ) frequency slightly varies around @xmath3 5 hz in observations # 35 ( band 1 ) , while no significant broad band components were fitted in observations # 6 and # 7 . in band 2 @xmath27 frequency shows a clear decreasing trend only in the last three observations ( @xmath34 @xmath3 4.41.6 hz ) , while does not show any clear trend in band 3 . @xmath27 rms decreases smoothly with time in band 1 ( from @xmath3 22@xmath23 to @xmath3 15@xmath23 ) , but it does not show the same trend in the other two energy bands . in band 2 we observe a clear decrease of @xmath27 rms only in observations # 56 ( from @xmath3 20@xmath23 to @xmath3 7@xmath23 ) and in band 3 it oscillates between 22% and 27% . + apart from the full energy , @xmath28 was fitted only in observation # 5 ( band 1 ) and # 6 ( band 13 ) , but it is significant just at low photon energy ( band 1 , # 6 ) . @xmath28 frequency and rms behavior in band 1 is mostly identical to band 0 . we fit the power spectra of the first 5 observations using propfluc ( id11 ; id12 ; ik13 ) . whereas original explorations of the model ( id11 ; id12 ) used computationally intensive monte carlo simulations , ik13 developed an exact analytic version of the model , allowing us for the first time to explore its capabilities systematically . we also investigate the relation between the values we obtained from the previously described phenomenological fitting of several lorentzians ( [ sec : obs ] and [ sec : res ] ) and the model physical parameters ( see table 2 in id12 for a summary and description of all the physical parameters ) . + ( fig . [ fig : simul]@xmath24 ) . ] propfluc ( [ sec : int ] ) parameterizes the flow surface density profile , which is required to calculate both the precession frequency and local viscous frequency , as a bending power law : @xmath43 where @xmath44 and @xmath45 is a break radius such that @xmath46 for @xmath47 and @xmath48 for @xmath49 , with the sharpness of the break controlled by the parameter @xmath50 ( fig . [ fig : den]b shows @xmath51 examples for different @xmath45 values ) . here , @xmath52 is the average mass accretion rate over the course of a single observation and @xmath53 is a dimensionless normalization constant . throughout this paper , we employ the convention that @xmath54 is radius expressed in units of @xmath55 . the surface density drop off at the bending wave radius , @xmath45 , is due to the torque created by the radial dependence of lt precession ( @xmath56 @xmath57 @xmath58 ) ; i.e. essentially the inner regions try to precess quicker than the outer regions . outside @xmath45 , bending waves ( pressure waves ) strongly couple the flow together but inside @xmath45 , material falls quickly towards the black hole ( lubow et al . 2002 ; fragile et al . the precession frequency of the flow is given by ( liu & melia 2002 ) : @xmath59 where @xmath60 is the keplerian orbital frequency and @xmath61 is the point particle lense - thirring ( lt ) precession frequency at @xmath62 ( id11 ) . the bending wave radius carries information about the shape of the inner flow because it is dependent on the scaleheight factor of the flow , @xmath63 : @xmath64 where @xmath65 is the dimensionless spin parameter . + if the mass is conserved on long timescales , the viscous frequency can be expressed as ( frank , king & raine 2002 ; id12 ) : @xmath66 propfluc assumes that the power spectrum of mass accretion rate fluctuations generated at @xmath62 is a zero - centered lorentzian cutting off at the viscous frequency . the model splits the flow up into rings and assumes that a constant fractional variability , @xmath67 , is generated per decade in radius . the resolution of the model is set by the number of rings per decade in radial extent , @xmath68 ; i.e. the interval between @xmath69 and @xmath70 is split into @xmath68 rings . consequently , the fractional variability in the mass accretion rate , @xmath71 , at each ring is @xmath72 , so that @xmath68 of these time series multiplied together has a fractional variability @xmath67 . the emitted luminosity from a ring at @xmath73 is then assumed to be @xmath74 , where the emissivity index @xmath75 is a model parameter , and the total emitted luminosity is simply the sum of the contributions from each ring . thus , the low frequency break in the power spectrum corresponds roughly to @xmath76 ( churazov , gilfanov & revnivtsev 2001 ; ingram & done 2010 ) . the high frequency break , however , does not correspond to @xmath77 because interference between radiation from different rings in the flow has a significant influence on the shape of the broad band noise at high frequencies ( id11 ; ik13 ) . the emissivity index also clearly affects the shape of the high frequency noise : increasing @xmath78 increases the amount of high frequency noise in the power spectrum . to better understand the relation between the phenomenological parameters derived from lorentzian power spectral fit characteristics and the physical parameters of propfluc , and to provide guidance in fitting this model to observed power spectra whose rough characteristics are known , we compute model power spectra with a nyquist frequency of 128 hz and vary one by one the main parameters . we fix the bh mass ( @xmath79 ) , the dimensionless spin parameter ( @xmath80 ) , the inner radius ( @xmath81 , so that @xmath82 ) , the bending wave radius ( @xmath83 , so that @xmath84 ) , and the emissivity index ( @xmath85 ) . we use a resolution @xmath86 and include a qpo with fixed width and rms . the results of the calculations are shown in fig . [ fig : simul ] . every plot specifies the parameter values , the asterisk in each panel denotes the value of that parameter used for all the other computations . [ fig : simul]@xmath18 shows that the centroid frequency of the broad band component ( hereafter @xmath87 ) decreases as @xmath53 increases , while the centroid frequency of the qpo ( hereafter @xmath88 ) does not change . this can be understood from equations [ eqn : fprec ] and [ eqn : fv ] . [ eqn : fv ] shows that increasing @xmath53 decreases @xmath89 but , since both integrals in eq . [ eqn : fprec ] contain the surface density , the constant @xmath53 cancels in the calculation of @xmath90 . in contrast , the _ shape _ of the surface density profile affects both the broad band noise and @xmath88 . [ eqn : fprec ] shows that the precession frequency of the entire flow is a surface density weighted average of @xmath91 ( the precession frequency of a test mass a distance @xmath62 from the black hole ) . therefore increasing the surface density at large @xmath62 slows down precession and increasing the surface density at small @xmath62 speeds up precession . [ eqn : fv ] and fig . [ fig : den]a b show that increasing the surface density in any region decreases the viscous frequency in that region and increasing the _ gradient _ of @xmath51 increases the _ range _ of frequencies at which the broad band noise contains significant power . [ fig : simul]@xmath24 shows that @xmath88 decreases with increasing @xmath45 . this is because @xmath45 governs where in the flow the surface density starts to drop off ( see fig . [ fig : den]b ) , and so increasing it weights the surface density towards the outer regions of the flow . since , as demonstrated in fig . [ fig : den ] , increasing @xmath45 slightly reduces the average surface density , this slightly increases the viscous frequency at the inner and outer rings in the flow , @xmath77 and @xmath76 . this causes @xmath92 to increase by a small amount ( which is difficult to see in fig . [ fig : simul]b because of the qpo ) , and its affect on the high frequency power is complicated by the emissivity and interference between radiation from different regions of the flow ( see ik13 ) . similar considerations are also valid for fig . [ fig : simul]c e . [ fig : simul]f shows @xmath88 increasing as @xmath4 decreases , roughly following the trend @xmath93 , where @xmath94 is a positive constant ( @xmath95 for the fiducial model parameters ) . this can be understood if we assume a constant surface density profile ( i.e. @xmath96 and @xmath45 is small ) and use the weak field approximation for lt precession , @xmath91 @xmath97 , to obtain a simplified version of eq . 2 in ingram , done , & fragile ( 2009 ) : @xmath98 } { r_o^{5/2 } r_i^{1/2 } [ 1-(r_i / r_o)^{5/2 } ] } \frac{c}{r_g}.\ ] ] we see that , for @xmath99 , the @xmath4 dependence of @xmath88 reduces to @xmath100 . we also see from this equation that increasing @xmath101 increases the qpo frequency , as demonstrated in fig . [ fig : simul]h . thus , the parameters which most affect the qpo frequency are @xmath4 ( @xmath3 quadratic ) and @xmath101 ( linear ) . as described in [ sec : modmod ] , the model is based on the inner accretion flow variability . for this reason , in order to avoid contamination from the disc , the best data choice for fitting would be observations in the high energy band ( @xmath3 1020 kev ) , as mentioned in id11 . because of the low count rate , we considered a wide 2.8720.20 kev band . the model assumes that all the variability is generated in the hot flow , so formally , in this scenario , the only effect of the disc is to suppress the variability amplitudes at lower energies by dilution . of course the possibility that the variability is generated in the disc and then propagates towards the inner flow affecting its emission , can not be excluded , but for our first explorative fit we just considered the simple scenario described above . we note that , using the spectral model _ tbabs * smedge * ( discbb + nthcomp + gauss ) _ ( wilms , allen & mccray 2000 ; zdziarski , johnson & magdziarz 1996 ; zycki , done & smith 1999 ; mitsuda et al . 1984 ) , we find that the disc contribution to the flux in band 1 increases from @xmath3 38@xmath23 in observation # 1 to @xmath3 61@xmath23 in observation # 7 ( it contributes significantly less in the other bands ) . since the rms in band 1 decreases from @xmath3 33@xmath23 to @xmath3 5@xmath23 for observations # 1 - 7 , our assumption that the disc is stable implies that the fractional rms of the flow in this band is @xmath3 53@xmath23 to @xmath3 13@xmath23 for observation # 1 - 7 respectively , which is reasonable . + since propfluc is not intended to explain the sims , we only fit the first 5 observations and leave a discussion of qualitative interpretation of observation # 6 and # 7 to [ sec : dis ] , in the absence of statistically acceptable fits . we fitted logarithmically binned data points in the frequency range 1/128128 hz , using the same resolution for data and model . we used @xmath68 = 15 for all the fits . compared to power spectra computed using higher resolution ( see [ sec : exp ] ) we did not observe any significant difference in @xmath35 . the difference produced by a larger number of rings is appreciable only in the higher frequency region of the power spectra , where our data error bars are large . we combined the qpo with the broad band variability by multiplication instead of addition mode ( the total flux is the product between the two types of variability instead of their sum , see ik13 ) . although our observations do not allow to differentiate between the multiplicative and the additive mode , the multiplicative mode is based on the more physically realistic scenario that the precession modulates the emission . because of the inner flow precession , the brightest part of the inner flow moves in and out of the observer s line of sight and angles to the line of sight vary , causing variations in the projected area ( ik13 ) . + for all the fits we fixed the parameters @xmath102 , @xmath103 , @xmath50 , the bending wave radius @xmath45 , the emissivity @xmath78 , the mass m , and the dimensionless spin parameter of the black hole @xmath65 . the free parameters are @xmath104 , the truncation radius @xmath105 , the fractional variability @xmath67 , the fundamental qpo width @xmath106 , and the rms of the fundamental and harmonic qpos . [ fig : propex ] shows the propfluc fit of the third observation , tab . [ tab : results ] shows the best fit parameters , the main qpo frequency @xmath90 , and the reduced @xmath107 for each of the 5 observations considered , and fig . [ fig : ingram ] shows the evolution of the physical parameters with time . + all parameter values show a clear trend . @xmath104 increases from @xmath3 3.5 to @xmath3 13 continuously . in the same way the truncation radius decreases from @xmath3 24 @xmath55 to @xmath3 10 @xmath55 , indicating an average truncation radius recession speed of about 2 km / h . the fractional variability increases from @xmath3 18@xmath23 to @xmath3 23@xmath23 in the first 3 observations and shows no significant change beyond . the fundamental qpo width increases continuously with time while its rms decreases from @xmath108 to @xmath109 . + error bars . ] as described by stiele et al . ( 2012 ) , the source follows the usual behavior of bhbs . using the state classification described in belloni ( 2010 ) , the source either is in the hs or in the hims during the first 5 observations . between observation # 5 and # 7 we observe rms dropping off ( from @xmath3 18@xmath23 to @xmath3 7@xmath23 ) simultaneously to the detection of radio emission ( miller - jones et al . 2011 ) , and in observation # 7 we identify a significant ( 4.7@xmath19 , single trial ) type - b qpo in the 2.8720.20 kev energy band . this shows the source entered the sims between observation # 5 and # 7 , a conclusion previously considered by stiele et al . ( 2012 ) , who however did not report the qpo in observation # 7 . + we used the profluc model ( id11 ; id12 ; ik13 ) for fitting the power spectra of the first 5 observations in the 2.87 - 20.20 kev band . as described in [ sec : int ] , the model assumes that the variability generated by each region of the inner flow peaks at the local viscous time scale . this is contrary to results of general relativistic magneto - hydrodynamic ( grmhd ) simulations , which display variability peaking approximately at the local _ orbital _ timescale ( dexter & fragile 2011 ; armitage & reynolds 2003 ) . these simulations , however , disagree with observations since bhbs display a high frequency break in their power spectra at @xmath110 hz , whereas simulations assuming a @xmath111 black hole exhibit variability up to a high frequency break of @xmath112 hz , @xmath31/3 the orbital frequency at @xmath113 ( e.g. , fig . 10 of dexter & fragile 2011 ) . the same discrepancy with simulations is seen for active galactic nuclei . for example , the power spectrum of the seyfert 1 galaxy ngc 4051 displays a break at @xmath114 hz ( vaughan et al . 2011 ) , whereas the orbital frequency at @xmath113 is @xmath115 hz ( assuming a black hole mass of @xmath116 ) . while it is clearly important that this inconsistency is addressed , we note that assuming the characteristic variability timescale to be orbital rather than viscous would require an inner flow radius of @xmath117 in order to fit the observed power spectra in this paper . this strongly implies that considering the viscous timescale is more appropriate for black holes , even in light of evidence that pulsars show variability on the orbital timescale ( revnivtsev et al . + in order to better explore the possibilities of the model , we ran a series of computations varying its physical parameters . the model parameter mainly affecting qpo frequency is the truncation radius @xmath4 , which sets the physical dimension of the precessing inner flow , responsible for the qpo generation . as can be seen in fig . [ fig : simul ] , most other parameters ( but not @xmath53 and @xmath67 ) also affect the qpo frequency . every parameter affects the broad band noise in a characteristic way . for example , varying the bending wave radius @xmath45 we observe evident correlated variations in peak frequency and slope of the high frequency tail of the broad band component , varying @xmath102 has almost no effect on the high frequency tail , but clearly changes its width and peak frequency , @xmath53 variations affect only the peak frequency . variations in black hole spin @xmath65 have no effect at all on the broad band component . it is clear that in order to constrain all physical parameters of the model , very good counting statistics are needed to clearly define the precise shape of the broad band noise . + in our fits we fixed @xmath103 , @xmath102 , @xmath50 , @xmath45 , the emission coefficient @xmath78 , black hole mass @xmath118 , and spin @xmath65 . fixing @xmath103 , @xmath102 , @xmath50 , and @xmath45 implies fixing the surface density profile throughout the rising phase of the outburst . id12 noticed that an evolution of @xmath45 is expected when the truncation radius decreases , because when the inner flow is illuminated by an increasing number of disc photons , its electron temperature drops and hence its scaleheight factor , @xmath63 , collapses . in our fit we fixed the bending wave radius because our spectra are too noisy above 10 hz to measure it independently , so we are not able to eliminate the degeneracy between @xmath4 and @xmath45 in determining the qpo frequency . for this reason we caution that the fit results in this work were obtained fixing the density profile of the inner flow ( except for @xmath4 ) , so that they must be interpreted with care . + the qpo frequency increase over observations # 15 corresponds to an @xmath4 decrease from @xmath3 24 to @xmath3 10 @xmath55 . from spectral analysis , stiele et al . ( 2012 ) report a constant truncation radius ( @xmath4 @xmath3 20 - 22 km ) throughout the whole outburst without specifying uncertainties . as described before , the profluc physical parameter mainly affecting the qpo frequency is @xmath4 , so that , in order to fit qpos during the rising phase of the outburst , the truncation radius has to vary during this phase . because of the data quality , rxte spectral range , and the limitations of the model used by stiele et al . ( 2012 ) ( see merloni et al . 2001 ) , the spectral estimation of the inner radius is of limited use in the comparison with the propfluc results . + the surface density normalization constant @xmath53 increases from @xmath3 3.5 to @xmath3 13.2 . for a given annulus in the inner flow , @xmath53 is proportional to surface density divided by mass accretion rate ( eq . [ eqn : sigma ] ) . during the rising phase of the outburst the mass accretion rate increases with time , so the @xmath53 increase means that the surface density increases faster than the mass accretion rate , i.e. matter is accumulating in the flow during this phase of the outburst . this is consistent with the results of id12 on xte j1550 - 564 . + the fractional variability @xmath67 shows a linear increasing trend in the first 3 observations and holds almost stable ( @xmath3 23@xmath23 ) in the other observations . the fractional variability does not give us any detailed information about the physical mechanisms producing the variability , but it quantifies the turbulent nature of the accreting material per radial decade . id12 show that @xmath67 increases continuously decreasing truncation radius , but we do not observe the same trend over all the observations , possibly because in our fit we fixed the bending wave radius . + as described , timing properties change in observations # 6 and # 7 , compared to observations # 15 , with simultaneous radio emission . the changes are more abrupt at higher photon energies ( fig . [ fig : rmsall ] ) . the rms decreases from @xmath3 18@xmath23 to @xmath3 7@xmath23 and @xmath32 frequency decreases as well , breaking the monotonically increasing trend of the first 5 observations . observations # 6 and # 7 also show a prominent low frequency broad band component ( @xmath119 ) that is not understandable in terms of the 2-component power spectra produced by profluc , which is the reason why we applied profluc only to the first 5 observations . belloni et al . ( 1997 ) consider emptying and replenishing cycles of the inner accretion disc , caused by viscous thermal instability , to explain variability on timescale of tens of minutes in grs 1915 + 105 . feroci et al . ( 1999 ) suggest material ejection to explain both spectral and timing properties of this same source , also in view of the correlation between the innermost disc temperature and the qpo frequency observed by muno et al . similarly , assuming the truncation radius reaches its smallest value at maximum luminosity ( and so maximum accretion rate ) , the @xmath90 decrease observed in our data can be explained by the depletion of inner flow material between @xmath5 and a certain radius @xmath120 ( @xmath121 ) simultaneously to the radio emission . this scenario is shown schematically in fig . [ fig : jetm ] . this depletion can be caused by either ejection or increased accretion between @xmath5 and @xmath120 . as a consequence , the surface density between @xmath5 and @xmath120 drops off , so that the high frequencies ( corresponding to smaller radii ) contribute less to the qpo frequency ( eq . [ eqn : fprec ] ) and @xmath67 ( so the noise level ) decreases . + the low frequency broad band component @xmath119 might be explained in this scenario as the result of mass accretion rate fluctuations propagating from the disc towards the inner flow . we analyzed the evolution of maxi j1543 - 564 during its 2011 outburst identifying the transition between lhs / hims and sims , occurring between observation # 5 and # 6 . analyzing the source in different energy bands , we found that in this transition changes in rms are more evident at higher photon energy . using the mass accretion rate fluctuation / precessing flow model propfluc , we provided a physical interpretation of the first 5 observations in terms of truncation radius , fractional variability , mass accretion rate , and surface density evolution . we suggest that the source behavior in observation # 6 and # 7 , and so the transition between lhs and sims , might be caused by mass depletion in the innermost part of the accretion flow due to ejection and/or enhanced accretion associated with the simultaneous radio emission . this physical scenario is consistent with our timing analysis in different energy bands . frank , j. , king , a. , & raine , d. j. 2002 , accretion power in astrophysics , by juhan frank and andrew king and derek raine , pp . isbn 0521620538 . cambridge , uk : cambridge university press , february 2002 . , & & 2.874.90 & 1.07@xmath123 & 4.17@xmath124 & 17.97@xmath125 & 4.86 & 1.03 + & & 4.909.81 & 1.03@xmath126 & 3.42@xmath127 & 20.80@xmath128 & 6.36 & 1.07 + & & 9.8120.20 & 1.04@xmath129 & 8.38@xmath130 & 17.42@xmath131 & 4.21 & 0.81 + & & 2.8720.20 & 1.06@xmath129 & 5.26@xmath132 & 17.48@xmath133 & 7.05 & 0.88 + & & 2.874.90 & 1.72@xmath134 & 6.54@xmath135 & 13.37@xmath136 & 6.32 & 1.17 + & & 4.909.81 & 1.75@xmath137 & 10.91@xmath138 & 14.34@xmath139 & 8.04 & 0.87 + & & 9.8120.20 & 1.74@xmath129 & 6.46@xmath140 & 18.69@xmath141 & 6.35 & 0.83 + & & 2.8720.20 & 1.75@xmath142 & 7.81@xmath143 & 14.74@xmath144 & 11.17 & 0.91 + & & 2.874.90 & 2.97@xmath129 & 7.82@xmath145 & 11.25@xmath146 & 10.49 & 1.04 + & & 4.909.81 & 2.97@xmath142 & 9.51@xmath147 & 16.06@xmath148 & 18.93 & 1.04 + & & 9.8120.20 & 2.97@xmath149 & 11.06@xmath150 & 16.37@xmath151 & 12.18 & 0.88 + & & 2.8720.20 & 2.97@xmath142 & 8.60@xmath152 & 14.60@xmath153 & 25.14 & 1.22 + & & 2.874.90 & 4.39@xmath154 & 12.35@xmath155 & 8.37@xmath156 & 4.86 & 0.84 + & & 4.909.81 & 4.42@xmath157 & 9.18@xmath158 & 14.57@xmath159 & 7.30 & 0.89 + & & 9.8120.20 & 4.46@xmath160 & 7.00@xmath161 & 19.62@xmath162 & 6.14 & 1.00 + & & 2.8720.20 & 4.381@xmath129 & 11.27@xmath163 & 13.01@xmath164 & 13.18 & 0.91 + & & 2.874.90 & 5.67@xmath165 & 6.61@xmath166 & 7.35@xmath167 & 2.93 & 1.04 + & & 4.909.81 & 5.84@xmath168 & 8.48@xmath169 & 11.10@xmath170 & 5.50 & 1.07 + & & 9.8120.20 & 5.83@xmath171 & 6,52@xmath172 & 17.41@xmath173 & 3.57 & 0.91 + & & 2.8720.20 & 5.77@xmath157 & 7.98@xmath174 & 10.43@xmath175 & 8.53 & 0.80 + & & 2.874.90 & 6.52@xmath176 & 3.69@xmath177 & 5.51@xmath178 & 2.19 & 1.10 + & & 4.909.81 & 5.74@xmath179 & 4.96@xmath180 & 6.08@xmath181 & 2.15 & 1.08 + & & 9.8120.20 & 6.19@xmath182 & 4.45@xmath183 & 12.84@xmath184 & 2.40 & 1.08 + & & 2.8720.20 & 5.75@xmath185 & 1.81@xmath186 & 6.35@xmath187 & 2.49 & 1.08 + & & 2.874.90 & 4.69 & 9.00 & @xmath188 5.32 & - & - + & & 4.909.81 & 4.83@xmath189 & 10.38@xmath190 & 5.34@xmath191 & 3.97 & 0.84 + & & 9.8120.20 & 4.86@xmath192 & 10.49@xmath193 & 11.37@xmath194 & 4.00 & 1.21 + & & 2.8720.20 & 4.69@xmath195 & 9.00 & 4.76@xmath196 & 6.48 & 1.01 + & & 2.874.90 & 2.28 & 4.55 & @xmath188 6.48 & - & - + & & 4.909.81 & 2.28@xmath197 & 4.55@xmath198 & 13.05@xmath199 & 2.53 & 1.07 + & & 9.8120.20 & 2.99@xmath200 & 4.55 & @xmath188 22.34 & 14.73@xmath201 & 0.81 + & & 2.8720.20 & 2.17@xmath165 & 6.78 & 5.42@xmath202 & 1.29 & 0.88 + & & 2.874.90 & 4.03@xmath203 & 5.18@xmath204 & 10.31@xmath205 & 2.78 & 1.17 + & & 4.909.81 & 4.03 & 5.18 & @xmath188 12.69 & - & - + & & 9.8120.20 & - & - & - & - & + & & 2.8720.20 & 4.03 & 5.18 & @xmath188 8.00 & - & - + & & 2.874.90 & 5.87@xmath206 & 5.18@xmath204 & 6.24@xmath207 & 2.50 & 1.04 + & & 4.909.81 & 5.87@xmath206 & 7.29@xmath208 & 6.57@xmath209 & 3.23 & 1.04 + & & 9.8120.20 & - & - & - & - & - + & & 2.8720.20 & 5.85@xmath210 & 6.55@xmath211 & 6.16@xmath212 & 4.23 & 1.22 + & & 2.874.90 & 8.86 & 6.69 & @xmath188 9.50 & - & - + & & 4.909.81 & 8.86 & 5.88 & 6.87@xmath213 & 1.55 & 0.89 + & & 9.8120.20 & - & - & - & - & - + & & 2.8720.20 & 8.86@xmath214 & 6.69@xmath215 & 5.88@xmath216 & 2.55 & 0.91 + & & 2.874.90 & - & - & - & - & - + & & 4.909.81 & 3.30 & 0 & @xmath188 30.16 & - & - + & & 9.8120.20 & - & - & - & - & - + & & 2.8720.20 & 3.30@xmath217 & 0 & 19.55@xmath218 & 4.22 & 0.88 + & & 2.874.90 & - & - & - & - & - + & & 4.909.81 & 3.641@xmath219 & 0 & 22.12@xmath220 & 6.64 & 0.87 + & & 9.8120.20 & 7.881@xmath221 & 0 & 24.65@xmath222 & 3.31 & 0.83 + & & 2.8720.20 & 3.61@xmath223 & 0 & 21.38@xmath224 & 11.96 & 0.91 + & & 2.874.90 & 4.813@xmath225 & 0 & 18.71@xmath226 & 7.72 & 1.04 + & & 4.909.81 & 3.755@xmath227 & 0 & 21.25@xmath228 & 8.38 & 1.04 + & & 9.8120.20 & 4.261@xmath229 & 0 & 27.40@xmath230 & 8.09 & 0.88 + & & 2.8720.20 & 3.77@xmath231 & 0 & 19.98@xmath232 & 15.83 & 1.22 + & & 2.874.90 & 4.64@xmath225 & 0 & 15.14@xmath233 & 4.48 & 0.83 + & & 4.909.81 & 3.44@xmath234 & 0 & 20.28@xmath235 & 6.02 & 0.89 + & & 9.8120.20 & 1.71@xmath236 & 0 & 21.61@xmath237 & 4.38 & 1.00 + & & 2.8720.20 & 3.42@xmath238 & 0 & 18.19@xmath239 & 8.69 & 0.91 + & & 2.874.90 & 5.36@xmath240 & 0 & 13.18@xmath241 & 3.52 & 1.04 + & & 4.909.81 & 4.36@xmath242 & 0 & 20.28@xmath243 & 8.11 & 1.07 + & & 9.8120.20 & 3.54@xmath244 & 0 & 25.89@xmath245 & 4.39 & 0.91 + & & 2.8720.20 & 3.69@xmath246 & 0 & 16.90@xmath247 & 8.53 & 0.80 + & & 2.874.90 & - & - & - & - & - + & & 4.909.81 & 2.50@xmath248 & 1.69@xmath249 & 6.94@xmath250 & 3.48 & 1.08 + & & 9.8120.20 & - & - & - & - & - + & & 2.8720.20 & 3.08@xmath251 & 0 & 8.56@xmath252 & 3.10 & 1.08 + & & 2.874.90 & 1.55 & 0 & @xmath188 5.03 & - & - + & & 4.909.81 & 1.55@xmath253 & 0 & 6.68@xmath254 & 3.32 & 0.84 + & & 9.8120.20 & - & - & - & - & - + & & 2.8720.20 & - & - & - & - & - + & & 2.874.90 & 0.06@xmath255 & 0 & 3.27@xmath256 & 2.15 & 1.04 + & & 4.909.81 & 0.06 & 0.13 & @xmath188 5.07 & - & - + & & 9.8120.20 & - & - & - & - & - + & & 2.8720.20 & 0.057@xmath257 & 0.13@xmath258 & 3.54@xmath259 & 3.04 & 0.80 + & & 2.874.90 & 0.19@xmath260 & 0 & 4.76@xmath261 & 4.49 & 1.10 + & & 4.909.81 & 0.26@xmath262 & 0 & 4.62@xmath263 & 2.86 & 1.08 + & & 9.8120.20 & 0.52@xmath264 & 0 & 10.47@xmath265 & 2.64 & 1.08 + & & 2.8720.20 & 0.13@xmath266 & 0 & 4.42@xmath267 & 5.00 & 1.08 + & & 2.874.90 & - & - & - & - & - + & & 4.909.81 & - & - & - & - & - + & & 9.8120.20 & - & - & - & - & - + & & 2.8720.20 & 0.27@xmath268 & 0 & 2.49@xmath269 & 2.23 & 0.91 +	we present a spectral and timing analysis of the black hole candidate maxi j1543 - 564 during its 2011 outburst . as shown in previous work , the source follows the standard evolution of a black hole outburst . during the rising phase of the outburst we detect an abrupt change in timing behavior associated with the occurrence of a type - b quasi - periodic oscillation ( qpo ) . this qpo and the simultaneously detected radio emission mark the transition between hard and soft intermediate state . we fit power spectra from the rising phase of the outburst using the recently proposed model propfluc . this assumes a truncated disc / hot inner flow geometry , with mass accretion rate fluctuations propagating through a precessing inner flow . we link the propfluc physical parameters to the phenomenological multi - lorentzian fit parameters . the physical parameter dominating the qpo frequency is the truncation radius , while broad band noise characteristics are also influenced by the radial surface density and emissivity profiles of the flow . in the outburst rise we found that the truncation radius decreases from @xmath0 to @xmath1 , and the surface density increases faster than the mass accretion rate , as previously reported for xte j1550 - 564 . two soft intermediate state observations could not be fitted with propfluc , and we suggest that they are coincident with the ejection of material from the inner regions of the flow in a jet or accretion of these regions into the bh horizon , explaining the drop in qpo frequency and suppression of broad band variability preferentially at high energy bands coincident with a radio flare . = -0.5 cm [ firstpage ] x - rays : binaries accretion , accretion discs - x - rays : individual ( maxi j1543 - 564 )
most stars in the galactic disk are formed in dense molecular gas within giant molecular clouds ( gmcs ) as members of clusters ( lada & lada 2003 ) . thus , clusters must play a critical role in the origins of some of the most fundamental properties of the galactic stellar population . in addition , such clusters consist of stars of various masses , and the most massive stars exist at the centers of clusters ( e.g. , raboud & mermilliod 1998 ) . these features suggest that the physical environment in the cluster - forming region would vary from area to area because the mass of formed stars should be closely related to the physical environment . however , clusters do not always include massive stars . some clusters contain only intermediate - mass and/or low - mass stars in low - mass star - forming regions ( e.g. , tachihara et al . 2002 ) . based on these features , it appears that although the conditions of massive star formation easily satisfy the physical conditions for cluster formation , the conditions of cluster formation do not necessarily provide the physical conditions for massive star formation . the clusters appear to be continuously forming in the galactic disk and the direct study of their physical conditions and the processes leading to their formation is basically possible . however , discovering clusters at an early evolutionary stage via optical observations is difficult because such clusters are completely embedded in dense gas and dust . therefore , observation by infrared instruments is necessary to study cluster formation . recently , many young embedded clusters within molecular clouds have been discovered by infrared observations ( e.g. , bica et al . 2003 ) . on the basis of these observations , to investigate the physical conditions in the dense gas that forms rich clusters , dense molecular gas surveys were carried out on such young clusters that include massive stars , revealing that clusters are formed in gas systems with a size scale of a cluster ( @xmath1 pc ) in gmcs ( e.g. , lada 1992 ; lada & lada 2003 ; saito et al . 2007 ) and these systems are usually called `` clumps . '' however , a few very compact gas systems having a size scale of @xmath5 pc similar to typical ( cold ) cores in low - mass star - forming regions ( e.g. , onishi et al . 2002 ; umemoto et al . 2002 ) have been discovered with very high - angular - resolution observations using an interferometer on young clusters that include massive ( proto)stars ( e.g. , churchwell et al . 1992 ; olmi et al . these systems have high temperatures ( @xmath6 k ) , a high h@xmath4 density ( @xmath7@xmath8 @xmath9 ) , and a high luminosity ( @xmath10 @xmath11 ) . these results indicate that these systems called `` hot cores , '' would include massive protostar(s ) or massive young star(s ) . generally , not a large number of stars form in a core even in cluster - forming regions ( e.g. , van der tak et al . 2005 ) . hence , many cores must form in one clump to create a cluster with a large number of stars . if this multiple - core system is formed by the fragmentation of the inner structure of the clump , a close relationship should exist between the physical parameters of the clump and the characteristics of the formed cluster . saito et al . ( 2007 ) carried out molecular line observations using the nobeyama 45-m telescope directed toward the clumps in several cluster - forming regions and suggested that a good relationship indeed exists between the average h@xmath4 density of the clumps and the stellar number density of formed clusters in these clumps . in addition , they found a good correlation between the line width of the clumps and the maximum mass of formed stars . saito et al . ( 2006 ) investigated the cores in three cluster - forming regions and found two types of cores in a single clump . the first type was a core with a large mass and a large turbulent motion , and the second type was a core with a small mass and a small turbulent motion . they found that massive ( proto)stars are associated with the cores with a large mass and large turbulent motion . this suggests a theoretical dependence between the mass of the star formed in a core and the turbulent motion in the core ( e.g. , mckee & tan 2003 ) . according to the suggestion by mckee & tan ( 2003 ) , the results of saito et al . ( 2006 ) indicate that massive and low - mass stars can be formed from one clump . such observations suggest that to understand the mechanism of cluster formation , it is important to elucidate the relationships between the physical properties of cores and physical condition of the clump that includes the cores , and between the mass of the star formed in a core and the physical condition of the core . if the cores in a clump are formed by gravitational instability as the theory suggests , a strong connection must exist between the spatial distribution of the cores in the clump and the inner structure of the clump . however , no detailed relationships between the cores and the clumps have been demonstrated . in addition , although young clusters have a tendency to concentrate massive stars in the center , whether massive cores with large turbulent motion are formed in the center of the clump remains unclear . therefore , we must observe several clumps to obtain detailed physical parameters of both clumps and cores and reveal the relationships between the cores and the clumps themselves . to detect the cores in the clump , we must uncover the internal motion and the distribution of the h@xmath4 column density in the clump . thus , we need to carry out molecular line observations suitable for revealing velocity structures of the dense gas in the clump . such observations are important to find a protocluster that consists of multiple cores a `` core cluster . '' we carried out c@xmath0o observations toward nine iras point sources with young clusters containing massive stars . the c@xmath0o line is a good candidate for tracing a large dynamic range in mass and column density in the dense gas because this line has a thin optical depth and a low critical density . in general , the c@xmath0o line can not be used to probe high - density regions ( @xmath12 @xmath9 ) in starless cores because co molecules are depleted onto dust grains in such a cold region with a temperature lower than the co sublimation temperature of @xmath13 k ( e.g. , caselli et al . 1999 ; bacmann et al . however , previous co observations have revealed that co lines trace high - density regions ( @xmath12 @xmath9 ) well with high temperatures of @xmath14 k ( e.g. , momose et al . 1998 ; schier et al . 2004 ) . therefore , if the temperatures of high - density regions ( @xmath12 @xmath9 ) are higher than @xmath13 k , c@xmath0o lines can be used to probe these regions . typically , the regions where massive stars are formed in gmcs maintain high temperatures of @xmath14 k , and thus c@xmath0o molecules would not be depleted in such regions . in fact , the structures identified by the interferometric observations toward massive star - forming regions have a high average density of @xmath15 @xmath9 ( e.g. , hunter et al . 2004 ; saito et al . therefore , c@xmath0o observations can be used to reveal the distribution , mass , and kinetic motion of the dense gas in cluster - forming regions if this line is optically thin enough . the identification of the star - forming and starless cores requires observation with high - angular resolution of several arc - seconds because massive star - forming regions are generally located at a distance of several kpc . however , interferometric observation of c@xmath0o cores in massive star - forming regions has rarely been performed . to find dense cores in the clumps with cluster formation , we carried out c@xmath0o molecular line observations with high - angular resolution toward nine cluster - forming regions using the nobeyama millimeter array ( nma ) . the physical conditions of the clumps in these regions were already derived by saito et al . results of the interferometric observations toward three of these nine regions were already reported in saito et al . ( 2006 ) . here we present the results of the observations toward the other six regions . we present our millimeter interferometric observations in @xmath16 2 . in @xmath16 3 , the observational results are presented for millimeter continuum and c@xmath0o molecular line . we discuss the implications of our results for the physical condition of the dense cores in @xmath16 4 . in @xmath16 5 , we summarize the main results of our study . to compare the physical conditions in the dense gas around young clusters containing young massive stars with the characteristics of the clusters , we need to select targets at similar evolutionary stages and distances and with similar bolometric luminosities , because the observed physical parameters depend on both the characteristics of formed stars ( e.g. , saito et al . 2001 ) and the spatial resolutions . in addition , the evolution of an embedded stellar cluster is sensitively coupled to the evolution of its surrounding gas ( e.g. , lada & lada 2003 ) . carpenter et al . ( 1990 ) , zinchenko et al . ( 1994 , 1997 , 1998 ) , and sridharan et al . ( 2002 ) carried out observations with low - angular resolution ( @xmath17 ) toward luminous iras point sources associated with h@xmath4o / ch@xmath18oh masers and/or compact hii regions in the northern sky and identified 93 massive clumps with a size of @xmath19 pc . h@xmath4o / ch@xmath18oh masers and/or compact hii regions serve as the indicators of very young massive stellar objects in massive star - forming regions . we selected 14 iras point sources from these clumps on condition that the luminosity of the iras source is several times @xmath20 @xmath11 , the distance is @xmath212 4 kpc , and the source is associated with both the indicators of very young massive stellar objects and nir clusters . we observed nine of these where the physical parameters of the dense clumps were already derived by single - dish observations with high - angular resolution of @xmath22 ( saito et al . 2007 ) . in this paper , we report the observational results obtained toward six objects except for three objects that were already reported by saito et al . these iras sources are associated with the indicators of very young massive stellar objects : ( 1 ) ( u)c hii region(s ) for iras05375 + 3540 ( e.g. , felli et al . 2006 ) , iras06117 + 1350 ( turner & terzian 1985 ) , iras19442 + 2427 ( e.g. , barsony 1989 ) , and iras19446 + 2505 ( e.g. , gmez et al . ( 2 ) very weak centimeter continuum sources for iras06056 + 2131 , and iras06061 + 2151 ( kurtz et al . this indicates that these massive ( proto)stars are at a similar evolutional stage within several times 10@xmath23 yr . table 1 lists the position , distance , and luminosity of each source . the observations were carried out from november 2002 through may 2005 with the six - element nobeyama millimeter array ( nma ) in the c@xmath0o ( @xmath24@xmath25 ; 109.782 ghz ) line , and 98 ghz and 110 ghz continuum . the 8-h tracks were obtained in the most compact ( designated as d ) and intermediate ( designated as c ) configurations . the projected antenna baseline length ranged from 13 m to 163 m , yielding the typical resolution of @xmath26 when natural weighting is applied to all visibility data . the resolution of each observation is given in table 2 . at 110 ghz , the primary beam of the 10-m telescope is about @xmath27 . we used the sis receivers , whose system noise temperatures in single sideband were @xmath28 k ( including the atmosphere toward the zenith ) . digital correlators uwbc and fx were configured for 1024 mhz and 32 mhz bandwidths with 128 and 1024 channels per baseline , respectively , yielding an equivalent resolution of 44 km s@xmath2 and 0.1 km s@xmath2 at 110 ghz , respectively . the uwbc can provide both the upper sideband ( 110 ghz ) and the lower sideband data ( 98 ghz ) . the gain calibrations were carried out for each upper and lower sideband every 10 - 15 min by observing quasars 0528 + 134 , 0552 + 398 , and 2023 + 336 . the bandpass calibration was performed with 3c454.3 and 3c84 . uranus was observed at 3 mm to estimate the absolute flux scale . the uncertainties in the flux scale are about 15% . the visibility data were calibrated using the nro software package uvproc2 , and final maps were made using the clean algorithm of the astronomical image processing system ( aips ) by the national radio astronomy observatory ( nrao ) . the 3-mm continuum maps are constructed from line - free channels of a 44 km s@xmath2 wide window . the uncertainties of the position in the maps are about @xmath29 . the noise levels of the natural - weighted maps are summarized in table 2 . we defined a compact structure , a core , from the distribution of c@xmath0o intensity to estimate physical properties of the dense condensations . we used the following procedures , similar to those adopted by saito et al . ( 2007 ) : \(1 ) to identify the velocity components in the observed fields , we checked the c@xmath0o molecular spectrum . if the c@xmath0o spectrum has more than two line features and the separations among their radial velocities are larger than 0.5 km s@xmath2 , each line feature was attributed to different velocity components . \(2 ) we made a c@xmath0o integrated intensity map for each velocity component . \(3 ) we searched for the i - th intensity peak ( peak - i ) outside previously defined cores ( i.e. , core[1 ] , core[2 ] , ... , core[i-1 ] ) in the c@xmath0o integrated intensity map . \(4 ) we drew a contour at the half - level of the intensity value at peak - i and defined the region enclosed by the half - intensity contour as a core[i ] . \(5 ) in cases when sub - peaks exist within the core[i ] , if the intensity at a sub - peak is larger than 6 @xmath30 above the hollow between it and peak - i , we excluded the region around the sub - peak by splitting the core[i ] along the hollow . \(6 ) in cases when previously defined cores ( i.e. , core[1 ] , ... , core[i-1 ] ) exist within the boundary of the core[i ] , if the intensity at peak - i is larger than 6 @xmath30 above the hollow between peak - i and the peaks of the previously defined cores , we excluded the region around the peaks of the previously defined cores by splitting the core[i ] along the hollow . otherwise , the core[i ] was canceled . \(7 ) we searched for the next ( [ i+1]-th ) intensity peak outside the cores . \(8 ) we repeated the procedures after procedure ( 4 ) until the value of the intensity of the local peak went below the 6 @xmath30 noise level . we identified 171 c@xmath0o molecular cores using these procedures . the observed properties of the c@xmath0o cores , such as the absolute peak temperature , @xmath31 , the radial velocity , @xmath32 , the fwhm line width , @xmath33 , at the peak position , and the composite line width , @xmath34 , of the core are summarized in table 3 . here , the composite line width is the fwhm line width of the composite spectrum made by averaging all spectra within the core . the typical uncertainties of the temperature , radial velocity , and line width in the gaussian fits are 0.1 k , 0.1 km s@xmath2 , and 0.1 km s@xmath2 , respectively . we next defined the millimeter continuum sources ( mcss ) from the continuum images to estimate the physical properties of the very dense regions and to identify the positions of the massive ( proto)stars . therefore , we identified mcss except the known hii regions from the 98 ghz and 110 ghz continuum maps using procedures ( 3)(8 ) adopted by the c@xmath0o core definition . we also stipulated that the mcss were detected at both the 98 ghz and the 110 ghz continuum bands . as a result , we identified 13 mcss and detected six known ( u)c hii regions . the observed parameters of these sources are summarized in table 4 . the goal of this study was to clarify physical processes of both cluster formation and massive star formation by investigating the indispensable physical characteristics of the dense gas that forms young clusters . in this paper , we defined a clump as a fundamental structure in which clusters are formed and defined a core as a fundamental structure in which individual stars in the cluster are formed . previous studies ( e.g. , umemoto et al . 2002 ; saito et al . 2006 , 2007 ) indicated that the clumps have a size scale of @xmath35 0.3 pc and the cores have a size scale of @xmath5 pc . the physical quantities of the dense gas in the cluster - forming regions , including massive star formations , are derived from the c@xmath0o - line emission . generally , the c@xmath0o molecule is chemically stable and the optical depth of c@xmath0o ( @xmath24@xmath25 ) is very thin . thus , this line is very useful for estimating the h@xmath4 column density . here , we calculate the c@xmath0o optical depth and the h@xmath4 column density under the local thermodynamic equilibrium ( lte ) condition with the excitation temperature , @xmath36 , estimated from the @xmath37co peak temperature of @xmath31(@xmath37co ) = 31 56 k at the position of each maser source ( barsony 1989 ; wouterloot & brand 1989 ; zinchenko et al . 1998 ; white & fridlund 1992 ) . we assume that the excitation temperature of each core is the same as that at the position of the maser source in each region . in these cases , the excitation temperatures , @xmath36 , were estimated to be 35 60 k. with these excitation temperatures , the c@xmath0o optical depth , @xmath38(c@xmath0o ) , and c@xmath0o column density , @xmath39(c@xmath0o ) , were estimated using @xmath40^{-1}\biggm\},\end{aligned}\ ] ] and @xmath41}\biggm\ } ( { \rm cm^{-2}})\end{aligned}\ ] ] ( e.g. , rohlfs & wilson 2004 ) , where @xmath31(c@xmath0o ) is the c@xmath0o radiation temperature in kelvins . we calculated the h@xmath4 column density , @xmath39(h@xmath4 ) , assuming that the c@xmath0o abundance ratio is @xmath42 ( frerking et al . the c@xmath0o optical depth and h@xmath4 column density at the intensity peak of the cores range from 0.02 to 0.19 and from @xmath43 to @xmath44 @xmath45 , respectively . from these results , we found that the optical depth of the c@xmath0o line is very thin . first , we estimated the line width and the radius of the cores . we regard the line width of the composite line profile as the line width of the core . the line width of the cores ranges from 0.43 to 3.33 km s@xmath2 with a mean line width of 1.08 km s@xmath2 . we estimated the radius of the cores assuming that the cores are spherical . we defined the radius of the core , @xmath46 , as the deconvolved radius and calculated it as @xmath47 where s and beam area are the area inside the core and the area of the synthesized beam , respectively . the radius ranges from @xmath48 to @xmath49 pc , with a mean radius of @xmath50 pc . we selected the cores with the ratio of the deconvolved radius to the observed radius of @xmath51 to select the cores resolved by the present observations . note that we call the radius calculated from ( s/@xmath52)@xmath53 the `` observed radius . '' hereafter , we use only the cores with high reliability for discussions involving the radius . next , we estimated the lte mass , @xmath54 , and the average h@xmath4 density , @xmath55(h@xmath4)@xmath56 , of the cores . the lte mass of the cores was derived by integrating the h@xmath4 column density over the core . we assumed a mean molecular weight per one h@xmath4 molecule of 2.8 by taking into account a relative helium abundance of 25% by mass . the average h@xmath4 density of the core is derived by assuming both a spherical shape and a uniform density and estimated by dividing the mass by a core volume , @xmath57 . the lte mass and the average h@xmath4 density of the cores range from 0.5 to 54.1 @xmath3 and @xmath58 to @xmath59 @xmath9 , respectively . finally , we estimated the virial mass , @xmath60 , of the cores . the virial mass is one of the key quantities by which we can evaluate the equilibrium state of the cores . the virial mass of the core is derived using the following equation , assuming isothermal , spherical , and uniform density distribution with no external pressure or magnetic forces : @xmath61 where @xmath62 is the gravitational constant . the virial mass of the cores ranges from 0.7 to 143.4 @xmath3 . these physical properties are summarized in table 5 . here , we distinguish starless cores and star - forming cores to investigate the star - formation activity in c@xmath0o cores in cluster - forming regions . we compared the cores with near infrared sources ( 2mass sources ) , mcss , and centimeter continuum sources ( ccss ) as indicators of young stellar objects to reveal star formation in the cores . in particular , we selected the 2mass sources detected at more than or equal to two bands with @xmath63 mag among the sources associated with the clumps ( saito et al . 20007 ) because the sources embedded in the dense gas must have a large extinction such that the typical peak h@xmath4 column density of the c@xmath0o cores is more than @xmath64 @xmath45 . note that the 2mass sources identified by saito et al . ( 2007 ) include young stellar objects ( ysos ) , which are the class i iii type sources , and background sources . here , we define an association as when an intensity peak of one object is located inside the boundary of the other object . note that the cores associated with only mcss are possibly very massive and dense starless cores . first , we compared 13 mcss with ccss and 2mass sources . three of these are associated with both 2mass sources and ccss , and four of 13 mcss are associated only with 2mass sources . only one of 13 mcss , mcs b for iras06056 + 2131 , is associated only with ccss and is seen as a dark lane in the @xmath65-band image in figures 2(b ) and ( c ) . the flux density of the ccss associated with mcss corresponds to that of the hii region formed by a b23 star or of the ionized stellar wind formed by a b01 star ( e.g. , panagia 1973 ; felli & panagia 1981 ) . therefore , the ccss associated with mcss would be caused by the activity of massive ysos . we also derived spectral indices of mcss from 98 ghz and 110 ghz data in table 6 . the spectral index of 8 mcss with 2mass sources and/or ccss ranges from 2.2 to 4.6 . this indicates that the millimeter continuum emission from these mcss is subject to thermal dust radiation . thus , the four mcss with ccss indicate the existence of young massive ( proto)stars . three other mcss only associated with 2mass sources except for mcs a for iras19442 + 2427 have a strong intensity peak in millimeter continuum emission , and 2mass sources associated with these mcss have the brightness corresponding to an early b type or earlier star . therefore these three mcss would include early b type or earlier ( proto)stars . indeed , two of these mcss , mcs a for iras05375 + 3540 and mcs a for iras06117 + 1350 , have a high luminosity of @xmath66 @xmath11 ( corresponding to a b34 star ) and @xmath67 @xmath11 ( corresponding to a o9.5 star ) , respectively ( eiroa et al . 1994 ; felli et al . although the 2mass source associated with mcs a for iras19442 + 2427 has a brightness corresponding to an early b star , this mcs has no strong intensity peak in millimeter continuum emission and the total integrated continuum emission from the mcs is very weak . in addition , although this mcs has a large h@xmath4 column density ( @xmath68 @xmath45 ) , the associated bright 2mass source has no large extinction ( @xmath69 ) thus , this 2mass source would be located in front of the mcs . however , four of five mcss with neither 2mass sources nor ccss , mcss a , b , and c for iras19446 + 2505 and mcs c for iras19442 + 2427 , have a flux density of 110 ghz similar to or smaller than that of 98 ghz , and the spectral indices of these four mcss are to be in the range between @xmath70 and @xmath71 in table 6 . in particular , three mcss for iras19446 + 2505 are located at the edge of the optical hii region , s88b . thus , we regard these four mcss as consisting of only free - free emission . note that mcs b for iras19446 + 2505 with a large negative index of @xmath70 is more extended diffusely than the other mcs . generally , it is probable that the missing flux of the 110 ghz observation is larger than that of the 98 ghz observation for a diffuse extended source because the minimum uv length of the 110 ghz observation is longer than that of the 98 ghz observation . therefore , the mcs b would have such a negative index . thus , we regard this mcs as a source consisting of only free - free emission . the other mcs with neither 2mass sources nor ccss , mcs b for iras19442 + 2427 , has a strong intensity peak and a large spectral index of @xmath72 , which is similar to that of the other massive - star - forming mcss . hence , in this paper , we regard this mcs as a source with massive protostars . next , we compared 171 c@xmath0o cores with mcss , 2mass sources , and ccss . eight of the cores are associated with both mcss and 2mass sources and two of 171 cores are associated only with mcss . in addition , 19 of 171 cores are associated only with 2mass sources . from this result , all mcss except for four mcss classified as f.f . ( see table 6 ) are associated with c@xmath0o cores . however , cores n and r for iras19442 + 2427 are associated with same 2mass source and same mcs , mcs a. the intensity distribution of the mcs corresponds to the part with the high h@xmath4 column density of these cores in figures 5(b ) and ( d ) . the mcs would be associated with high - density part of these cores . therefore , cores n and r for iras19442 + 2427 with mcs a , which consist of cold dust emission , would be starless cores rather than cores with massive ysos . from these results , we regard 27 cores associated with 2mass sources and/or mcss except for cores n and r for iras19442 + 2427 as star - forming cores . in particular , we regard eight star - forming cores associated with mcss that include massive protostars as massive - star - forming cores . therefore , we identified 27 star - forming cores , including eight massive - star - forming cores . these results are summarized in table 5 . note that the nir source associated with these cores is not necessarily a ( proto)star itself , and the possibility exists that this nir emission is a reflection nebula of one or more central heating source(s ) . in addition , it may be probable that some of the cores cataloged as starless are associated with mid - ir sources . generally , the 98 and 110 ghz continuum emissions consist of free - free emission and/or thermal dust emission . first , we derived the flux of thermal dust component in mcss with ccss . the flux density of ccss are much smaller than the total flux of mcss at the millimeter wavelength as shown in table 4 . the spectral indices at 100 ghz of uc hii regions and ionized winds are usually @xmath73 and @xmath74 ( e.g. , panagia & felli 1975 ) , respectively . we assumed that these ccss consist of the free - free emission with a spectral index of 0 from an optically thin hii region or ionized winds , and we estimated the flux of thermal dust component in mcss by subtracting the contribution of the free - free emission from the total flux of mcss . next , we estimate the physical properties of the mcss . the deconvolved radii of the mcss , @xmath75 , are calculated using equation ( 3 ) . the estimated radius of mcss ranges from 0.018 to 0.110 pc . in particular , the radius of the mcss associated with c@xmath0o cores ranges from 0.018 to 0.055 pc and is smaller than that of the associated cores ( @xmath75/@xmath76 ) . in addition , assuming an optically thin dust thermal emission at 100 ghz , we estimated the mass of mcss using @xmath77 where @xmath78 is the gas and dust mass of an mcs , @xmath79 is the total flux of the mcs , @xmath80(@xmath81 ) is the planck function with the assumed dust temperature , @xmath82 , and @xmath83 is the dust opacity per gram with @xmath84 @xmath85 g@xmath2 ( preibisch et al . 1993 ) , which assumes a standard gas - to - dust mass ratio of 100 . here , we assume that the dust opacity index , @xmath86 , of the dust in mcss is 1.5 and that the dust temperature is equal to the temperature of the gas . first , we need to determine the temperatures of mcss to estimate the mass of mcss . the temperature of mcss a and b for iras05375 + 3540 were estimated by felli et al . ( 2004 ) using ch@xmath87cn lines , and we use this temperature as the dust temperature of the mcs . next , because no observation has been carried out toward the other mcss using molecular lines , which trace the gas temperatures such as ch@xmath18cn and nh@xmath18 , we estimate the temperature of mcss with massive star formation to be @xmath88 @xmath89 k by applying the stefan boltzmann law using the luminosity of ( proto)stars associated with the mcss . if we estimate the temperature of mcss a and b for iras05375 + 3540 using the same method , the temperature of @xmath90 k is obtained , which is similar to the temperature derived by molecular line described above . concerning mcs a for iras19442 + 2427 without massive star formation , we use the excitation temperature estimated from the @xmath37co peak temperature as the dust temperature . we estimate the mass of mcss using these dust temperatures and the mass of the mcss ranges from @xmath91 to @xmath92 @xmath3 . we estimate the average h@xmath4 density of the mcss using the radius and the mass of the mcss , and the average h@xmath4 density ranges from 1.0 @xmath93 to 8.1 @xmath93 @xmath9 . the integrated intensity maps of c@xmath0o as well as various kinds of objects related to star formation are given in figures 1 6 . for each region , the distribution of c@xmath0o clumps , which were taken from saito et al . ( 2007 ) , iras sources , h@xmath4o masers , hii regions , and the primary beam area of the present nma observation are given in figure ( a ) . in addition , the physical properties of the clumps are summarized in table 7 . the integrated intensity distribution of c@xmath0o in the clumps is compared with the 98 ghz or 110 ghz continuum image in figure ( b ) . in figure ( c ) , the c@xmath0o distribution is overlaid on the 2mass @xmath65-band image . the distribution of c@xmath0o cores , ccss , and mcss are given in figure ( d ) . the strong peak in the continuum image indicate massive ( proto)stars embedded in dense gas and the @xmath65-band sources associated with cores indicate young stars or protostars . in addition , although each missing flux in the nma primary beam area is roughly estimated to be @xmath94% using the single dish observation data ( saito et al . 2007 ) , the missing flux toward the present c@xmath0o cores is small to be @xmath90% . this suggests that the missing flux of the present observations would be caused by the structure more extended than the size of the one measured with the present observations . as the interferometer filters out the extended emission and c@xmath0o observations can detect the dense gas , the present observations would accurately detect compact cores in a clump . in the following sections , we present descriptions of the individual regions . in figure 1 , we identified 21 c@xmath0o cores and two mcss in three clumps , clumps b , c , and d identified by saito et al . in addition , we detected free - free emission from a part of the hii region , s235 a ( e.g. , felli et al . 2006 ) , by the present continuum observation . c@xmath0o core j is associated with mcs a and a bright 2mass source . although this 2mass source is associated with the intensity peak of mcs a , the intensity peak of core j has an offset by @xmath26 from this 2mass source . c@xmath0o core h is associated with mcs b , a bright 2mass source , and a ccs ( felli et al . 2006 ) . in figure 2 , we identified 19 c@xmath0o cores and two mcss in clump a identified by saito et al . c@xmath0o core e is associated with mcs a and a bright 2mass source . although the intensity peak of mcs a is associated with this 2mass source , the intensity peak of core e has an offset by @xmath95 from this 2mass source . c@xmath0o core j is associated with mcs b and two ccss and the intensity peak of this core is associated with a dark lane in the @xmath65-band image . in figure 3 , we identified 24 c@xmath0o cores and two mcss in clump b identified by saito et al . c@xmath0o core o is associated with mcs b , two ccss , and two bright 2mass sources . the intensity peak of this core is associated with one of these ccs and is located between these 2mass sources . c@xmath0o core p is associated with mcs a , a ccs , and a bright 2mass source . the intensity peak of core p has an offset by @xmath96 from this ccs . in addition , this ccs has a larger flux density ( @xmath97 mjy ) than that ( @xmath98 mjy ) of the ccss associated with core o. in figure 4 , we identified 28 c@xmath0o cores and one mcs in two clumps , clumps c and d identified by saito et al . we also detected the free - free emission from a part of the hii region , s269 ( e.g. , turner & terzian 1985 ) , by the present continuum observation . the intensity peak of c@xmath0o core m is located at both an intensity peak of mcs a and the position of a bright 2mass source , irs2w . according to the luminosity of irs2w ( @xmath99 ) , c@xmath0o core m may have an embedded massive ( proto)star . in figure 5 , we identified 38 c@xmath0o cores , three mcss in three clumps , clumps a , b , and c identified by saito et al . in addition , we detected the free - free emission from both the core and the extended structure of hii region s87 ( e.g. , barsony 1989 ) . two velocity components , 21 km s@xmath2 and 24 km s@xmath2 components , exist in this region , and it has been suggested that the cluster , gl2454 , was formed by the cloud cloud collision of these components ( saito et al . 2007 ) . c@xmath0o cores , cores r , u , and ac , around the core structure of hii region s87 have a large mass ( @xmath100 @xmath3 ) and a large line width ( @xmath101 km s@xmath2 ) . c@xmath0o core ah is associated with mcs b , which has a strong intensity peak and a large spectral index , although this core is not associated with bright 2mass sources . this indicates that the core could harbor an embedded massive yso(s ) . in addition , the position of the intensity peak of this core in 20.7 km s@xmath2 component map has changed from that in 21.4 km s@xmath2 component map in figure 5(d ) , and mcs b is located between the intensity peak in each component map . this indicates that core ah has a velocity gradient and could have a disklike system embedded in it . we suggest that this embedded massive yso(s ) in mcs b ( see @xmath16 3.2 ) would be a massive protostar . in figure 6 , we identified 41 c@xmath0o cores , three mcss in two clumps , clumps b and c identified by saito et al . in addition , we detected millimeter continuum emission , which corresponds to two compact hii regions s88b1 and s88b2 ( e.g. , gmez et al . 1995 ) . the distribution of the c@xmath0o emission is consistent with that of the dark region in the @xmath65 band . c@xmath0o cores , cores q and w , around the compact hii region s88b1 have a large mass ( @xmath102 ) and a large line width ( @xmath103 km s@xmath2 ) . three mcss exist at the boundary between the dense molecular gas and the optical hii region s88b . this indicates that the surface of the dense molecular gas would be ionized by the uv radiation from the massive stars in gl2455 . saito et al . ( 2007 ) identified 39 clumps with a size scale of @xmath1 pc in cluster - forming regions including massive star formation , and discussed their physical conditions by analyzing the correlations between the physical quantities of the clumps . they found that the average h@xmath4 density of the clumps increases with the line width and the clumps are self - gravitationally bound because they have a virial mass comparable to a gas mass . here , we extend their analysis to the inner structure of the clumps ( i.e. , dense cores ) and attempt to clarify the physical condition of the dense cores that are the direct site of each star formation in the cluster - forming region . first , we analyzed the relationships among the physical parameters of the cores . we used 171 c@xmath0o cores identified in this study and 28 c@xmath0o cores identified by saito et al . the cores identified by saito et al . ( 2006 ) were classified into three massive - star - forming cores , seven star - forming cores , and 18 starless cores using the classification adopted in this study . these cores exist in 18 dense clumps identified by saito et al . next we discuss the physical conditions of the dense cores inside the clumps in cluster - forming regions including massive star formation . the line width of molecular line emissions provides us with internal kinetic properties such as thermal motion and turbulence . several studies have suggested that a correlation exists between the line width and the radius of the core / cloud in the form of a single power - law relationship @xmath104 ; @xmath33 and @xmath105 are a characteristic line width and the radius of the core / cloud ( e.g. , larson 1981 ) . in addition , the index @xmath106 has recently been shown to depend on the environment ( e.g. , caselli & myers 1995 ; saito et al . 2006 ) . the left panel of figure 7 shows the relationship between the line width and the radius of the 181 cores with high reliability ( see @xmath16 3.1.2 ) identified by this work and saito et al . ( 2006 ) and 18 clumps identified by saito et al . note the other 18 cores are those with a large error in the radius in @xmath16 3.1.2 . although the line width of the clumps is in a narrow range of 1.5 to 3.2 km s@xmath2 , the line width of the cores is in a much wider range of 0.43 to 3.33 km s@xmath2 . in contrast , the range of the radius of the cores is relatively small . such features are seen even in the individual regions in the right panels of figure 7 . we investigated the index @xmath106 of the relationship between the line width and the radius of the core and the parental clumps . we estimated this relationship , @xmath107 = @xmath108 ( @xmath109)@xmath110 , for all cores . the index @xmath106 of this relationship ranges from 0.06 to 0.89 and the long and the short dashed lines in figure 7 indicate the line width radius relationships with the minimum and maximum index @xmath106 , respectively , using a mean line width and a mean radius of the clumps in the plots . our results show that the relationship between the line width and the radius does not follow a single power - law relation . this trend agrees with the result of saito et al . ( 2006 ) , who used a smaller number of cores and clumps . this indicates that the degree of the dissipation of the turbulent motion varies even within a single clump . as discussed below in @xmath16 4.1.3 , this may be due to the difference in the initial h@xmath4 density or the external pressure of the core . we postulate that the difference in the dissipation of the turbulent motion depends on the region where the core is formed in the clump , considering that the clump generally has a density structure . generally , a core has to be gravitationally bound to result in star formation . thus , a core with a large line width needs to have a large mass or high external pressure for star formation to occur . here , we check the virial condition of the cores to investigate the stability of the system . figure 8 shows correlations among the lte mass , @xmath111 , and the virial mass , @xmath60 , for 181 cores . most of the cores have a larger virial mass than the lte mass , and no clear difference is seen between star - forming cores and starless cores . the cores have a virial ratio ( the ratio of the virial mass to the lte mass ) of @xmath112 , although several massive - star - forming cores have somewhat larger virial ratios of @xmath113 . generally , more extended structures with star formation ( such as clumps ) have a similar virial mass to a lte mass ( fontani et al . 2002 ; saito et al . the main reason for the difference in the virial ratio between the clumps and the cores is the elapsed time after the structures were formed . the virial mass of the clump just after formation would be also larger than the lte mass and the clump would be gravitationally bound by the external pressure . the virial mass of the clump will later become similar to the gas mass because the density structure and the internal kinetic motion of the clump change and cluster formation is expected to occur in the clump . therefore , almost all of the clump would be virialized as far as we observe the clumps with cluster formation . however , many cores would also exist just after formation in the clump . thus , many `` young cores , '' which have a larger virial mass than the gas mass , could be detected by the present observation . in addition , the virial mass of the cores , except for massive - star - forming cores , is approximately fitted by a single power - law function by @xmath114 where c.c . is a correlation coefficient . the coefficient of this relationship is close to unity , indicating that in the virial equation ( see below ) , the difference in the gravitational and kinetic energy terms is almost proportional to the gravitational energy terms . to strike a good balance between this difference and the external pressure and for more massive cores to achieve the equilibrium state , the external pressure on the surface of the cores must be larger . most of the cores require external pressure to maintain the structures . here , we estimate the external pressure required to bind the present cores gravitationally . the virial equation is expressed by neglecting the magnetic fields as follows : @xmath115 where @xmath116 and @xmath117 are the kinetic and gravitational energy of the core , respectively , @xmath118 is the external pressure on the core surface , and @xmath62 and @xmath30 are the gravitational constant and the velocity dispersion , @xmath34/(@xmath119 ) . the external pressure required to bind the core gravitationally , @xmath120 , can be estimated using these equations with @xmath121=@xmath25 . the required pressure of the cores with a virial ratio of @xmath19 ranges from @xmath122 to @xmath123 k cm@xmath2 . next , we investigated the external pressure on the core surface . since the external pressure is difficult to estimate , we used the average pressure in a clump ( average clump pressure ) @xmath125 instead , although these values generally differ from each other . the average clump pressure was estimated using both the line width , @xmath126 , and the average h@xmath4 density , @xmath127 , of the clump . we estimated the average clump pressure using the parameters of the clumps identified by saito et al . ( 2007 ) in table 7 and found that the average clump pressure ranges from @xmath128 to @xmath129 k cm@xmath2 . figure 9 shows the required pressure of cores plotted against the projected distance from the center of the clump to the core ( core distance ) , @xmath130 . here , the required pressure and the core distance are normalized to the average clump pressure and the radius of the clump , @xmath131 , respectively . we found that the required pressure for most of the cores , except for massive - star - forming cores , is smaller than the average clump pressure . therefore , our findings suggest that most of the cores are likely to be gravitationally bound by the external pressure . finally , we investigated how the required pressure of cores changes with the position in a clump . in figure 9 , the cores with a large @xmath132/@xmath133 are concentrated in the center of the clump . therefore , the nearer the cores are to the center of the clump , the higher the external pressure on the core surface is expected to be . when the h@xmath4 density structure of the clump follows @xmath55(h@xmath4)@xmath134 , the pressure in the clump ( clump pressure ) , @xmath135 , is estimated as a function of @xmath130 by @xmath135/@xmath133 = ( @xmath136/3)@xmath137/@xmath131)@xmath138 . here , we assumed that the turbulent velocity is constant within the clump . for comparison , the distributions of the clump pressure for @xmath139 = 1.0 , 1.5 , and 2.0 are indicated in figure 9 by the dotted , solid , and dashed lines , respectively . these lines are similar to the upper limit of the @xmath132/@xmath133@xmath130/@xmath131 plot , which indicates that almost all cores would be gravitationally bound by the external pressure if the h@xmath4 density structure in the clump has a relation of @xmath55(h@xmath4)@xmath134 . in addition , these results suggest that the external pressure of the core surface would depend on the h@xmath4 density structure in the clump . here , we investigated the relationship between the lte mass and the line width of the core . figure 10 shows that the lte mass of the cores increases in proportion to the square of the line width , although the dispersion of the plot is large . this tendency agrees with the result of @xmath16 4.1.2 : from equations ( 4 ) and ( 6 ) , we obtain the relationship @xmath140 considering that the radius of the cores is almost constant . from this tendency , dense gas with a large kinetic motion in a clump seems to be necessary to form massive cores because the lte mass of the cores increases with the line width . figure 10 also indicates that no clear difference exists between star - forming and starless cores , although several massive - star - forming cores ( black circles ) have somewhat larger line widths than the other cores with a similar mass . for example , the average line widths of massive - star - forming cores and the starless cores ( massive starless cores ) with a mass @xmath141 @xmath3 are calculated as 2.8 km s@xmath2 and 2.0 km s@xmath2 , respectively . the main cause of the difference in the line width is supposed to be the effect of star formation . if massive stars in massive - star - forming cores have formed from massive starless cores with a similar mass of @xmath142 @xmath3 , the increase in momentum due to star formation is roughly estimated to be @xmath143 @xmath3 km s@xmath2 . the estimated increase in momentum is much smaller than the outflow momentum supplied by massive protostar(s ) . for example , the total outflow momentum supplied by massive protostar(s ) with a luminosity of @xmath144 @xmath11 is estimated to be @xmath145 @xmath3 km s@xmath2 ( zhang et al . 2005 ) . this indicates that at most only a tenth the total outflow momentum is injected to the present core , and almost all of the outflow momentum is likely to be supplied into the intercore gas or outside the clump . in other words , massive - star - forming cores should have a large line width even prior to star formation . therefore , we suggest that although star formation activity tends to enhance the line width of the cores , massive starless cores should form from a dense gas with a large kinetic motion . finally , considering that the range of the radius of the cores is small , the relationship between the mass and the line width of the cores suggests that the average h@xmath4 density of the cores would increase with the line width of the cores . according to this result and the discussion on the external pressure in @xmath16 4.1.2 , a small dissipation of the turbulent motion is sufficient for a core formed in the center of the clump with the h@xmath4 density structure to be gravitationally bound because the initial condition of this core necessarily has a high h@xmath4 density and a high external pressure . in this case , the index @xmath106 of the line width radius relationship from the clump to the core would become very small ( @xmath146 ) . however , a core formed at the edge of the clump must undergo a large dissipation of turbulent motion to be gravitationally bound , and the index @xmath106 would become large ( @xmath147 ) . in other words , the physical parameters of the core would depend on the place where the core is formed in the clump . this relationship is discussed in detail in @xmath16 4.2.2 . here , we discuss the relationships between the mass of the star formed in the core and the physical parameters of the core and between the physical parameters and the positions of the cores in the clump . generally , the mass of a star is expected to depend on the physical properties of the parental core in a clump . thus , the characteristics of the stellar member of the cluster would depend on the characteristics of the cores in the clump . in addition , saito et al . ( 2007 ) found that the maximum mass of the stars associated with the clump increases with the line width of the clump and that a good correlation exists between the stellar number density of the cluster and the average h@xmath4 density of the clump . these results indicate the importance of revealing the relationships among the clump , the cores , and young ( proto)star(s ) to investigate both cluster formation and massive star formation . typical massive - star - forming cores have a higher h@xmath4 density , larger line width , and higher temperature than the low - mass star - forming cores ( e.g. , churchwell et al . 1992 ; olmi et al . although many observations with such tendencies have been reported , the relationship between the mass of the star and the physical properties of the parental core has never been quantitatively investigated . the main reason is that the discovery of the cores forming stars with various masses in the same region has been very difficult . however , typical stellar clusters have many stars with various masses , and stars formed in star - forming cores described here are expected to have a wide range of masses . therefore , we estimated the stellar masses of the stars using the same method as saito et al . ( 2007 ) . for stars whose luminosities have been estimated or stars with centimeter - continuum sources , we estimated their spectral types using these parameters . we estimated the spectral types of other stars from the @xmath148/@xmath149 color magnitude diagrams using the 2mass catalog . figure 11 shows the relationships between the mass of ( proto)stars associated with cores , @xmath150 , and the physical parameters , lte mass and line width , of the cores . a tendency exists for the mass of stars to increase with these parameters . these results indicate that the mass of the star would depend on the mass ( or the h@xmath4 density ) and internal kinetic motion of the parental core . generally , high average h@xmath4 density and large internal kinetic motion would be expected to lead to a high mass accretion rate ( e.g. , mckee & tan 2003 ) . although the plot of the relationship between the mass of the stars and the mass of the cores has a large scattering , the plot can roughly be expressed by the relationship of @xmath150 = @xmath54 , which is indicated by the solid line . from this relationship , the star - formation efficiency ( sfe ) is roughly constant at @xmath151% using the equation of sfe = @xmath150/(@xmath150 + @xmath54 ) . note that the mass of the core is expected to decrease with the age of the protostar(s ) . based on the above relationships and the mass line width relationship in @xmath16 4.1.3 , the magnitude of internal kinetic energy of the core would determine the mass of the formed stars . in addition , the core must have existed in the clump at least during the lifetime of the clump for star formation to occur therein . it is therefore necessary for the core to be gravitationally bound . from these results , we suggest that the necessary and sufficient condition to form a higher mass star is that the core has a larger internal kinetic motion and is gravitationally bound . in this case , it is not necessary for the core to be self - gravitationally bound if high external pressure is brought to bear on the core surface . because the internal kinetic motion basically decreases with time , a structure such as a molecular cloud much larger than a core or a clump must previously have had a large internal kinetic motion and must be gravitationally bound for massive star formation to occur . thus , gmcs and clouds with an external effect , for example , hii regions and supernova remnants , would be good environments for massive star formation because such clouds would have a large external pressure and be supplied with large kinetic energy . here , we discuss the spatial distribution of the physical parameters of the cores in the clump . in @xmath16 4.1 , we found that dense gas with a large kinetic motion in a clump seems to be necessary to form massive cores , and the external pressure on the surface of cores must be large for massive cores to achieve an equilibrium state . in addition , cores with large external pressure are concentrated in the center of the clump . thus , the physical properties of the core are expected to depend on the physical condition of the region where the core is formed in the clump . first , we examined the relationship between the physical parameters , line width and lte mass , of the cores and the core distance ( see @xmath16 4.1.2 ) . figure 12(a ) shows the masses of the cores plotted against the core distance . although no clear tendency is seen in this plot , the upper limit of the core mass would gradually decrease with the core distance . if the average h@xmath4 density of the cores is proportional to the h@xmath4 density in the clump having a density structure of @xmath152 , at the region where the core was formed , the core mass can be expressed by the relationship @xmath153 where @xmath154 and @xmath155 are the mean molecular weight per one h@xmath4 molecule and the hydrogen mass , respectively , and @xmath156 and c are the h@xmath4 density in the clump at the distance of @xmath131 from the center of the clump and the proportional coefficient , respectively . generally , the h@xmath4 density structure of a clump is suggested to have a relationship with ( @xmath130/@xmath157 ( @xmath1582.0 ; e.g. , van der tak et al . 2000 ) . here , if a clump has a mass of 300 @xmath3 and a radius of 0.3 pc and @xmath46 and c are 0.03 pc and 50 , which are typical respective values for the present sample , the mass distance relationships with @xmath139 = 1.0 and 2.0 are indicated by the solid and dashed lines in figure 12(a ) , respectively . as shown in figure 12(a ) , the distribution of the upper limit of the core mass is roughly similar to the relationships expressed by the two lines . from this result , the mass ( or the average h@xmath4 density ) of the core is suggested to depend on the h@xmath4 density of the region where the core is formed in the clump , and the distribution of the core mass is suggested to be controlled by the h@xmath4 density structure in the clump . figure 12(b ) shows the ratio of the line width for the core and the clump plotted against the core distance . the starless cores with a high line width ratio ( @xmath159 ) are concentrated in the center of the clump , and the upper limit of this ratio would gradually decrease with the core distance . the reason for this feature would be the change in both the core mass and the external pressure on the core surface . here , if we assume the cores are gravitationally bound by the clump pressure , we can estimate the maximum line width of the cores using the virial equation ( see @xmath16 4.1.2 ) . if the clump with the physical parameters adopted in the above discussion is gravitationally bound , a line width of the clump is 2.2 km s@xmath2 . in this case , the ratio of the maximum line width of the cores to the line width of the clump with @xmath139 = 1.0 and 2.0 is indicated by the solid and dashed lines , respectively , in figure 12(b ) . as shown in figure 12(b ) , the distribution of the upper limit of the line width ratio is similar to the relationships expressed by the two lines . as we have seen in @xmath16 4.1.2 and the above - mentioned mass distance relationship , the external pressure on the core surface and the core mass are likely to depend on the core distance due to the h@xmath4 density structure in the clump . therefore , to bind gravitationally , the internal kinetic motion of the core would also necessarily depend on the core distance . from the discussions in this subsection , we suggest that the central region of the clump , which has a density structure of @xmath152 , is the environment in which the core with a large internal kinetic motion and a high average h@xmath4 density is easily formed and that the internal kinetic motion and the average h@xmath4 density of the core would depend on the h@xmath4 density structure in the clump . in addition , according to the result in @xmath16 4.2.1 , the mass of a star formed in the core would decrease with the distance from the center of the clump . therefore , the physical characteristics of the cluster would depend on the h@xmath4 density structure , internal kinetic motion , and the mechanism of core formation in the clump . finally , we discuss the relationship between cluster formation and the dense gas structures , cores and clumps . as we have seen in @xmath16 4.1 and @xmath16 4.2 , many cores with a size of @xmath5 pc that are formed in a clump with a size of @xmath1 pc have a wide range of the line width and the mass , and the spatial distribution of the physical parameters of the cores would depend on the h@xmath4 density structure in the clump . in addition , the mass of the star would depend mainly on the internal kinetic motion of the parental core . the stellar number density of the cluster is mentioned as an important characteristic of a cluster . the stellar number density of the cluster is generally much higher than that of typical low - mass star - forming regions . in this regard , saito et al.(2007 ) found a good relationship between the number density of ysos associated with the clump and the average h@xmath4 density of the clump , and they suggested that it is necessary for a clump to have a high average h@xmath4 density for cluster formation to occur therein . this suggests that a relationship may exist between the number density of the cores in the clump and the average h@xmath4 density of the clump . first , we estimated the number density of the cores to investigate this relationship . figure 13(a ) shows the relationship between the number density of the cores in the clump , @xmath160 , and the average h@xmath4 density of the clump , @xmath161 . a good correlation exists between the number density of the cores and the average h@xmath4 density of the clump . this relationship is roughly expressed by the equation @xmath162 @xmath161@xmath163 and the index of this relationship is similar to the index ( @xmath103 ) of the relationship between the number density of the ysos and the average h@xmath4 density of the clump obtained by saito et al . this indicates that star formation in the clump is based on core formation in the clump . although these relations are not expressed by the mechanism of core formation using simple gravitational instability as proposed by saito et al . ( 2007 ) , we suggest that for core formation with a high number density of the cores , it is necessary for the clump to have a high average h@xmath4 density as an initial condition . next , we examine the relationship between the number of the cores and the ysos to reveal the timescale of the cores . figure 13(b ) shows the number of ysos , @xmath164 , associated with the clump , which were identified by saito et al . ( 2007 ) , plotted against the number of cores in the clump , @xmath165 . the numbers of the ysos are one to three times larger than those of the cores . considering that almost all of the ysos would be low - mass stars , the lifetime of the yso would be @xmath166 yr ( e.g. , lada 1999 ) . in this case , the lifetime of the core is roughly estimated to be @xmath167 yr if only one star is formed per core . in addition , we find 27 star - forming cores in this study and 10 in saito et al . ( 2006 ) , and the number corresponds to @xmath168% of all cores . this result indicates that the lifetime of the star - forming cores , which can be detected by our interferometric observation , is @xmath169 yr . this timescale is consistent with the fact that the typical mass of the cores associated with class ii stars , whose lifetime is @xmath170 yr ( e.g. , lada 1999 ) , was roughly estimated to be @xmath171 in low - mass star forming regions ( e.g. , tsukagoshi et al . 2007 ) because the detection limit of the mass of the cores is @xmath172 . therefore , most of the ysos in the clump are not associated with the cores with a mass of @xmath173 and the ysos associated with these cores are regarded as young protostars . finally , we look at the core - formation efficiency ( cfe ) of the clump . figure 13(c ) shows cfes ( @xmath174/@xmath175 , where @xmath54 and @xmath175 are the mass of the core in the clump and the mass of the clump , respectively ) plotted against the number density of cores in the clump . although no strong correlation is seen between the cfes and the number density of cores , the cfe gradually increases with the number density of cores in the clump . considering the relationship between the number density of cores and the average h@xmath4 density of the clump in figure 13(a ) , this feature indicates that the dense gas in a clump with a high average h@xmath4 density is efficiently converted into many cores . from the results that the cfe is @xmath17660% and that the lifetime of the core is @xmath167 yr in the above discussion , the conversion rate from the dense gas to cores per @xmath177 yr ( core - formation rate : cfr ) in the clump is estimated to be @xmath178 20%/@xmath7 yr . thus , in a clump with a low average h@xmath4 density , core formation with a low cfr ( @xmath35 a few % /@xmath7 yr ) could have occurred over a long time of @xmath179 yr if the cfr is always constant . however , only a part of the dense gas in the clump would be converted to cores because this timescale is longer than the typical lifetime of the gmc ( @xmath180 yr ; e.g. , blitz et al . 2007 ) . on the contrary , in a clump with a high average h@xmath4 density , core formation with a high cfr ( @xmath181%/@xmath7 yr ) could have occurred over a short time of @xmath182 yr . this indicates that most of the dense gas in the clump would be converted to cores because this timescale is shorter than the typical lifetime of the gmc . from the above considerations and the discussions in @xmath16 4.1 and 4.2 , we determined that the mechanism of cluster formation would strongly depend on the mechanism of core formation in a clump and that the mechanism of core formation would be controlled by both the h@xmath4 density structure and the kinetic motion in the clump . moreover , the characteristics of the cluster would be determined by the distribution of the physical parameters of cores in the clump . if a clump with a very high average h@xmath4 density ( @xmath183 @xmath9 ) , which has a size scale of @xmath1 pc , is formed in a gmc , this clump would be expected to form a cluster with a very high stellar number density very efficiently in a short time based on the discussion of the cfr . in addition , considering that the virial mass of the clump is similar to the lte mass ( saito et al . 2007 ) , the line width of the clump is estimated to be @xmath100 km s@xmath2 using the average h@xmath4 density and the radius . according to the line width radius relationship in @xmath16 4.1.1 and the stellar mass line width relationship in @xmath16 4.2.1 , the cluster formed in the clump would include very massive ( e.g. , o3 ) stars . thus , the clumps with a very high average h@xmath4 density ( @xmath183 @xmath9 ) would be expected to form a cluster with both a very stellar number density and very massive stars . in addition , if many clumps are formed in a gmc , star - forming regions within the gmc will include many dense clusters with very massive stars and will be extended with a size of @xmath145 pc because the typical size of the gmc is @xmath145 pc . indeed , very active star - forming regions occurring in external galaxies , for example , 30 doradus region in lmc , are extended with a size scale of @xmath184 pc and have many dense clusters , which contain many o3 stars , with a size scale of @xmath1 pc , for example , r136a core in 30 doradus . therefore , we suggest that the mechanism of cluster formation in such active star - forming regions would be basically similar to that in our galaxy , although the clumps in such active star - forming regions would have much higher h@xmath4 density and much larger kinetic motion than those in our galaxy . we made high - angular resolution observations with the nobeyama millimeter array ( nma ) of six cluster - forming regions including massive ( proto)stars , in the @xmath24@xmath25 transition of c@xmath0o molecular emission line and 98 ghz and 110 ghz continuum emission . we identified 171 c@xmath0o cores in 12 clumps toward six cluster - forming regions including massive ( proto)stars . based on our observations , we studied the correlation between several physical quantities of the c@xmath0o cores and clumps , and discussed the formation mechanism of star clusters . the main results of the present study are summarized as follows : \2 . we identified 13 millimeter continuum sources ( mcss ) . seven are associated with 2mass sources and/or centimeter continuum sources ( ccss ) . together with the one mcs that shows similar characteristics to the above seven mcss , these eight mcss would include massive protostars . the range of the mass of mcss is 2.7111 @xmath3 . in addition , we compared c@xmath0o cores with mcss , 2mass sources , and ccss . we identified 27 star - forming cores including eight massive - star - forming cores . some cores have various line widths in a single clump . the range of the index of the relationship between the line width and radius of the core and the parental clump differs from core to core . this implies that the degree of dissipation of the turbulent motion varies within a single clump . \4 . for most cores , the virial masses are about twice larger than the lte masses . this indicates that external pressure is necessary for the cores to be gravitationally bound . based on the virial analysis , the distribution of the external pressure required to bind the cores is likely to be comparable to that of the pressure in the clump with an h@xmath4 density structure of @xmath185 ( @xmath1862 ) . the mass of the cores increases with the line width . this indicates that the average h@xmath4 density of the cores would depend on the kinetic motion of the dense gas in the core because the range of the radius of the core is small . the mass of the star depends on the line width and the mass of the parental core . in addition , we found that the star - formation efficiency ( sfe ) is roughly constant at @xmath151% and that no clear dependence exists between the sfe and the mass of the formed stars . \7 . the mass , line width , and external pressure of the core decrease with distance from the center of the clump , and a relationship exists between these decreases and the h@xmath4 density structure of the clump . we suggest that the spatial distribution of physical parameters of the cores strongly depends on the h@xmath4 density structure in the clump . the number density of the cores and the number density of the young ( proto ) stars in the clump have a similar relationship to the average h@xmath4 density of the clump . in addition , the core - formation rate ( cfr ) of the clump is estimated to be @xmath17820% per @xmath187 yr and the cfr gradually increases with the average h@xmath4 density of the clump . we thank to the staff of the nobeyama radio observatory ( nro ) for both operating the millimeter array and helping us with data reduction . we thank to kazuyoshi sunada , tomofumi umemoto , and takeshi sakai and also to the referee for their valuable comments . this publication makes use of data produced by 2mass , which is a joint project of the university of massachusetts and infrared processing and analysis center funded by nasa and nsf . this study also made use of the simbad database . bacmann , a. , lefloch , b. , ceccarelli , c. , castets , a. , steinacker , j. & loinard , l. 2002 , , 389 , l6 barsony , m. 1989 , , 345 , 268 bertoldi , f. & mckee , c. f. 1992 , , 395 , 140 bica , e. , dutra , c. m. , & barbuy , b. 2003 , , 397 , 177 blitz , l. , fukui , y. , kawamura , a. , leroy , a. , mizuno , n. , & rosolowsky , e. 2007 , in protostars & planets v , ed . b. reipurth , d. jewitt & k. keil ( tucson : univ . arizona press ) , 951 carpenter , j. m. , snell , r. l. , & schloerb , f. p. 1990 , , 362 , 147 caselli , p. , & myers , p. c. 1995 , , 446 , 665 caselli , p. , walmsley , c. m. , tafalla , m. , dore , l. , & myers , p. c. 1999 , , 523 , l165 churchwell , e. , walmsley , c. m. , & wood , d. o. s. 1992 , , 253 , 541 crampton , d. , georgelin , y. m. , & georgelin , y. p. 1978 , , 66 , 1 eiroa , c. , casali , m. m. , miranda , l. f. , & ortiz , e. 1994 , , 290 , 599 evans , n. j. , ii , & blair , g. n. 1981 , , 246 , 394 felli , m. , & panagia , n.1981 , , 102 , 424 felli , m. , massi , f. , navarrini , a. , neri , r. , cesaroni , r. , & jenness , t. 2004 , , 420 , 553 felli , m. , massi , f. , robberto , m. , & cesaroni , r. 2006 , , 453 , 911 fontani , f. , cesaroni , r. , caselli , p. , & olmi , l. w. 2002 , , 389 , 603 frerking , m. a. , langer , w. d. , & wilson , r. w. 1982 , , 262 , 590 gmez , y. , garay , g. , & lizano , s. 1995 , , 453 , 727 hunter , t. r. , zhang , q. , & sridharan , t. k. 2004 , , 606 , 929 kennicutt , r. c. 1984 , , 287 , 116 kurtz , s. , churchwell , e. , & wood , d. o. s. 1994 , , 91 , 659 lada , e. a. 1992 , , 393 , l25 lada , c. j. 1999 , in the origin of stars and planetary systems , ed . c. j. lada & n. d. kylafis ( dordrecht : kluwer ) , 143 lada , c. j. , & lada , e. a. 2003 , , 41 , 57 larson , r. b. 1981 , , 194 , 809 malumuth , e. m. , & heap , s. e. 1994 , , 107 , 1054 mckee , c. f. , & tan , j. c. 2003 , , 585 , 850 moffat , a. f. j. , fitzgerald , m. p. , & jackson , p. d. 1979 , , 38 , 197 momose , m. , ohashi , n. , kawabe , r. , nakano , t. , & hayashi , m. 1998 , , 504 , 314 olmi , l. , cesaroni , r. , & walmsley , c. m. 1993 , , 276 , 489 onishi , t. , mizuno , a. , kawamura , a. , tachihara , k. , & fukui , y. 2002 , , 575 , 950 panagia , n. 1973 , , 78 , 929 panagia , n. , & felli , m. 1975 , 39 , 1 preibisch , th . , osseukopf , v. , yorke , h. w. , & henning , th . 1993 , , 279 , 577 raboud , d. , & mermilliod , j .- a . 1998 , , 333 , 897 . rohlfs , k. , & wilson , t. l. 2004 , in tools of radio astronomy ( berlin : springer ) saito , h. , mizuno , n. , moriguchi , y. , matunaga , k. , onishi , t. , mizuno , a. , & fukui , y. 2001 , , 53 , 1037 saito , h. , saito , m. , moriguchi , y. , & fukui , y. 2006 , , 58 , 343 saito , h. , saito , m. , sunada , k. , & yonekura , y. 2007 , , 659 , 459 schier , f. l. , jrgensen , l. k. , van dishoeck , e. f. , & blake , g. a. 2004 , , 418 , 185 sridharan , t. k. , beuther , h. , schilke , p. , menten , k. m. , & wyrowski , f. 2002 , , 566 , 931 tachihara , k. , onishi , t. , mizuno , a. , & fukui , y. 2002 , , 385 , 909 tukagoshi , t. 2007 preprint . turner , k. c. , & terzian , y. 1985 , , 90 , 59 umemoto , t. , kamazaki , t. , sunada , k. , kitamura , y. , & hasegawa , t. 2002 , in the proceeding of the iau 8th asian - pacific regional meeting , volume ii , ed . by s. ikeuchi , & t. hanawa , 229 van der tak , f. f. s. , van dishoeck , e. f. , evans , n. j. , ii , & blake , g. a. 2000 , , 537 , 283 van der tak , f. f. s. , tuthill , p. g. , & danchi , w. c. 2005 , , 431 , 993 wouterloot , j. g. a. , & brand , j. 1989 , , 80 , 149 white , g. j. , & fridlund , c. v. m. 1992 , , 266 , 452 zhang , q. , hunter , t. r. , brand , j. , sridharan , t. k. , cesaroni , r. , molinari , s. , wang , j. , & kramer , m. 2005 , , 625 , 864 zinchenko , i. , forsstroem , v. , lapinov , a. , & mattila , k. 1994 , , 288 , 601 zinchenko , i. , henning , th . , & schreyer , k. 1997 , , 124 , 385 zinchenko , i. , pirogov , l. , & toriseva , m. 1998 , , 133 , 337 cccccccccccc & & & & & + & & & distance & @xmath188 & cluster & ccs@xmath189 & ref.@xmath190 + & & & ( kpc ) & ( 10@xmath191 ) & & & + iras05375 + 3540 & 05 & 40 & 53.6 & 35 & 42 & 16 & 1.8 & 2.5 & gl5162 & s235a b & 1 + iras06056 + 2131 & 06 & 08 & 41.0 & 21 & 31 & 01 & 3.5 & 7.2 & gl6366s & weak emission & 2 + iras06061 + 2151 & 06 & 09 & 07.8 & 21 & 50 & 39 & 3.5 & 3.4 & gl5182 & weak emission & 2 + iras06117 + 1350 & 06 & 14 & 36.6 & 13 & 49 & 35 & 3.8 & 6.5 & gl902 & s269 & 3 + iras19442 + 2427 & 19 & 46 & 19.9 & 24 & 35 & 24 & 2.3 & 5.0 & gl2454 & s87 & 4 + iras19446 + 2505 & 19 & 46 & 47.0 & 25 & 12 & 43 & 2.0 & 13.0 & gl2455 & s88b1,s88b2 & 4 + cccccrc & & & & + object & @xmath106(j2000 ) & @xmath192(j2000 ) & line or continuum & & & + & ( h m s ) & ( @xmath193 @xmath194 @xmath195 ) & & ( @xmath196 ) & & + iras05375 + 3540 & 05 40 53.3 & 35 41 48.8 & 98 ghz cont . & @xmath197 & @xmath19834 & 0.79 + & & & 110 ghz cont . & @xmath199 & @xmath19836 & 0.97 + & & & c@xmath0o & @xmath199 & @xmath19865 & 110@xmath189 + iras06056 + 2131 & 06 08 40.4 & 21 31 00.6 & 98 ghz cont . & @xmath200 & @xmath19837 & 0.78 + & & & 110 ghz cont . & @xmath201 & @xmath19837 & 1.08 + & & & c@xmath0o & @xmath201 & @xmath19838 & 80@xmath189 + iras06061 + 2151 & 06 09 07.3 & 21 50 36.7 & 98 ghz cont . & @xmath202 & @xmath19839 & 0.61 + & & & 110 ghz cont . & @xmath203 & @xmath19839 & 0.93 + & & & c@xmath0o & @xmath203 & @xmath19829 & 78@xmath189 + iras06117 + 1350 & 06 14 47.2 & 13 49 39.3 & 98 ghz cont . & @xmath204 & @xmath19823 & 0.58 + & & & 110 ghz cont . & @xmath205 & @xmath19824 & 0.95 + & & & c@xmath0o & @xmath205 & @xmath19828 & 63@xmath189 + iras19442 + 2427 & 19 46 20.1 & 24 35 29.4 & 98 ghz cont . & @xmath206 & @xmath19840 & 0.77 + & & & 110 ghz cont . & @xmath207 & @xmath19840 & 0.80 + & & & c@xmath0o & @xmath207 & @xmath19841 & 80@xmath189 + iras19446 + 2505 & 19 46 47.0 & 25 12 43.2 & 98 ghz cont . & @xmath208 & @xmath19839 & 3.6 + & & & 110 ghz cont . & @xmath209 & @xmath19839 & 6.5 + & & & c@xmath0o & @xmath209 & @xmath19837 & 84@xmath189 + clcccccccc iras05375 + 3540 & core a & 05 40 51.3 & 35 41 40 & & 2.0 & @xmath19815.9 & 0.85 & 0.92 & 1.7 + & core b & 05 40 51.3 & 35 41 37 & & 3.0 & @xmath19816.3 & 0.87 & 0.96 & 2.6 + & core c & 05 40 52.3 & 35 41 39 & & 2.1 & @xmath19818.0 & 1.09 & 1.18 & 2.3 + & core d & 05 40 52.4 & 35 41 47 & & 2.5 & @xmath19817.3 & 0.89 & 0.85 & 2.2 + & core e & 05 40 52.4 & 35 42 02 & & 1.5 & @xmath19818.2 & 0.51 & 0.56 & 0.8 + & core f & 05 40 52.4 & 35 41 59 & & 1.3 & @xmath19818.7 & 0.49 & 0.53 & 0.7 + & core g & 05 40 52.8 & 35 41 18 & & 2.4 & @xmath19817.1 & 0.53 & 0.60 & 1.3 + & core h & 05 40 53.4 & 35 41 41 & & 2.3 & @xmath19819.4 & 1.14 & 1.11 & 2.7 + & core i & 05 40 53.4 & 35 41 29 & & 2.7 & @xmath19817.8 & 1.33 & 1.38 & 3.8 + & core j & 05 40 53.4 & 35 41 56 & & 5.1 & @xmath19818.2 & 1.21 & 1.31 & 6.6 + & core k & 05 40 53.6 & 35 41 17 & & 2.7 & @xmath19816.7 & 0.68 & 0.77 & 1.9 + & core l & 05 40 53.8 & 35 41 25 & & 2.0 & @xmath19816.8 & 0.70 & 0.75 & 1.5 + & core m & 05 40 53.9 & 35 41 40 & & 2.5 & @xmath19818.8 & 0.58 & 0.75 & 1.6 + & core n & 05 40 54.2 & 35 42 13 & & 2.5 & @xmath19816.6 & 0.87 & 0.87 & 2.4 + & core o & 05 40 54.2 & 35 41 49 & & 1.4 & @xmath19817.3 & 0.51 & 0.55 & 0.8 + & core p & 05 40 54.4 & 35 42 08 & & 4.0 & @xmath19816.3 & 0.89 & 0.94 & 3.9 + & core q & 05 40 54.5 & 35 41 55 & & 1.9 & @xmath19816.9 & 0.67 & 0.73 & 1.3 + & core r & 05 40 54.9 & 35 41 49 & & 1.4 & @xmath19815.8 & 1.23 & 1.21 & 1.8 + & core s & 05 40 55.0 & 35 41 58 & & 3.3 & @xmath19816.7 & 0.72 & 0.72 & 2.5 + & core t & 05 40 55.0 & 35 42 09 & & 5.9 & @xmath19815.9 & 0.70 & 0.77 & 4.4 + & core u & 05 40 55.3 & 35 42 02 & & 3.2 & @xmath19816.3 & 0.41 & 0.51 & 1.4 + & & & & & & & & & + iras06056 + 2131 & core a & 06 08 39.1 & 21 30 35 & & 4.7 & 2.8 & 0.87 & 0.97 & 4.1 + & core b & 06 08 39.8 & 21 30 58 & & 0.9 & 4.5 & 1.11 & 1.16 & 1.1 + & core c & 06 08 40.1 & 21 30 52 & & 2.9 & 2.6 & 1.36 & 1.62 & 4.0 + & core d & 06 08 40.1 & 21 30 38 & & 2.7 & 2.4 & 1.09 & 1.25 & 3.0 + & core e & 06 08 40.2 & 21 30 59 & & 3.3 & 2.4 & 2.81 & 3.00 & 9.5 + & core f & 06 08 40.2 & 21 31 18 & & 1.6 & 2.9 & 0.92 & 0.94 & 1.5 + & core g & 06 08 40.5 & 21 30 43 & & 2.7 & 1.6 & 1.55 & 1.67 & 4.5 + & core h & 06 08 40.6 & 21 31 17 & & 1.4 & 4.4 & 0.67 & 0.72 & 1.0 + & core i & 06 08 40.6 & 21 30 33 & & 1.3 & 0.9 & 0.90 & 0.84 & 1.2 + & core j & 06 08 40.8 & 21 31 07 & & 1.5 & 3.1 & 3.89 & 2.87 & 6.0 + & core k & 06 08 41.0 & 21 30 41 & & 1.8 & 3.6 & 1.40 & 1.57 & 2.6 + & core l & 06 08 41.1 & 21 30 55 & & 1.3 & 4.1 & 0.63 & 0.68 & 0.9 + & core m & 06 08 41.1 & 21 31 17 & & 1.4 & 4.3 & 0.68 & 0.78 & 1.0 + & core n & 06 08 41.6 & 21 30 50 & & 1.9 & 3.6 & 0.65 & 0.68 & 1.2 + & core o & 06 08 41.6 & 21 30 37 & & 3.2 & 3.6 & 1.14 & 1.31 & 3.6 + & core p & 06 08 41.7 & 21 31 00 & & 1.1 & 4.3 & 1.36 & 1.43 & 1.6 + & core q & 06 08 41.8 & 21 31 24 & & 2.3 & 3.8 & 0.68 & 0.72 & 1.5 + & core r & 06 08 42.2 & 21 30 59 & & 2.6 & 3.8 & 0.84 & 0.87 & 2.2 + & core s & 06 08 42.4 & 21 31 04 & & 2.6 & 3.6 & 1.11 & 1.14 & 3.0 + & & & & & & & & & + iras06061 + 2151 & core a & 06 09 05.3 & 21 50 32 & & 1.1 & @xmath1981.3 & 0.51 & 1.11 & 0.6 + & & & & & 2.0 & @xmath1980.6 & 0.53 & & 1.1 + & core b & 06 09 05.6 & 21 50 46 & & 1.8 & @xmath1981.1 & 0.84 & 0.92 & 1.5 + & core c & 06 09 05.8 & 21 50 47 & & 1.6 & @xmath1982.3 & 0.46 & 0.46 & 0.8 + & core d & 06 09 05.8 & 21 50 42 & & 1.2 & @xmath1983.4 & 1.04 & 1.13 & 1.3 + & core e & 06 09 05.9 & 21 50 32 & & 1.8 & @xmath1981.2 & 0.82 & 0.85 & 1.5 + & core f & 06 09 06.3 & 21 50 53 & & 1.5 & 0.3 & 0.82 & 0.87 & 1.3 + & core g & 06 09 06.4 & 21 50 24 & & 1.3 & @xmath1982.4 & 0.84 & 0.90 & 1.1 + & core h & 06 09 06.4 & 21 50 10 & & 1.9 & @xmath1980.9 & 0.99 & 1.02 & 1.9 + & core i & 06 09 06.4 & 21 50 55 & & 1.8 & 1.2 & 0.77 & 0.78 & 1.4 + & core j & 06 09 06.5 & 21 51 01 & & 1.5 & @xmath1981.7 & 0.61 & 0.58 & 1.0 + & core k & 06 09 06.6 & 21 50 11 & & 1.6 & @xmath1982.4 & 0.58 & 0.60 & 0.9 + & core l & 06 09 06.7 & 21 51 06 & & 1.9 & 2.3 & 0.84 & 0.85 & 1.6 + & core m & 06 09 06.9 & 21 50 31 & & 1.0 & @xmath1982.6 & 1.21 & 1.26 & 1.4 + & core n & 06 09 07.0 & 21 51 01 & & 1.8 & 0.7 & 0.53 & 0.61 & 0.9 + & core o & 06 09 07.0 & 21 50 41 & & 2.4 & @xmath1981.9 & 1.04 & 3.24 & 2.6 + & & & & & 0.9 & @xmath1980.7 & 0.85 & & 0.8 + & & & & & 1.8 & 0.3 & 1.19 & & 2.3 + & & & & & 0.6 & 1.7 & 0.68 & & 0.4 + & & & & & 0.5 & 2.6 & 0.60 & & 0.3 + & core p & 06 09 07.1 & 21 50 38 & & 1.3 & @xmath1981.8 & 1.25 & 2.73 & 1.6 + & & & & & 0.8 & 0.5 & 0.51 & & 0.6 + & core q & 06 09 07.2 & 21 50 54 & & 1.6 & 1.3 & 0.92 & 0.97 & 1.5 + & core r & 06 09 07.2 & 21 50 17 & & 1.5 & @xmath1982.2 & 0.72 & 0.77 & 1.1 + & core s & 06 09 07.6 & 21 50 19 & & 2.1 & 0.7 & 1.23 & 1.30 & 2.7 + & core t & 06 09 07.7 & 21 50 24 & & 2.4 & 0.0 & 1.28 & 1.31 & 3.0 + & core u & 06 09 08.1 & 21 50 11 & & 3.2 & 0.7 & 0.77 & 0.80 & 2.5 + & core v & 06 09 08.4 & 21 50 33 & & 1.3 & @xmath1980.5 & 0.99 & 1.16 & 1.3 + & core w & 06 09 08.6 & 21 50 52 & & 2.3 & @xmath1980.3 & 1.14 & 1.21 & 2.7 + & core x & 06 09 09.3 & 21 50 24 & & 2.1 & @xmath1980.3 & 1.26 & 1.31 & 2.9 + & & & & & & & & & + iras06117 + 1350 & core a & 06 14 35.6 & 13 49 28 & & 2.0 & 15.7 & 0.70 & 0.77 & 1.4 + & core b & 06 14 35.7 & 13 49 31 & & 1.8 & 21.4 & 0.94 & 1.01 & 1.8 + & core c & 06 14 36.0 & 13 49 58 & & 2.0 & 19.2 & 0.46 & 0.67 & 0.9 + & core d & 06 14 36.1 & 13 49 16 & & 1.5 & 19.4 & 0.60 & 0.70 & 0.9 + & core e & 06 14 36.2 & 13 49 14 & & 2.9 & 15.8 & 0.44 & 0.46 & 1.3 + & core f & 06 14 36.4 & 13 49 17 & & 2.6 & 19.6 & 0.44 & 0.46 & 1.1 + & core g & 06 14 36.5 & 13 49 22 & & 1.5 & 19.5 & 0.65 & 0.68 & 1.0 + & core h & 06 14 36.6 & 13 49 32 & & 1.0 & 15.8 & 0.67 & 0.75 & 0.7 + & core i & 06 14 36.7 & 13 49 28 & & 1.9 & 17.7 & 0.49 & 0.53 & 0.9 + & core j & 06 14 36.7 & 13 49 38 & & 1.0 & 19.6 & 1.45 & 1.54 & 1.5 + & core k & 06 14 36.8 & 13 49 16 & & 1.6 & 18.5 & 0.75 & 0.78 & 1.2 + & core l & 06 14 36.9 & 13 49 14 & & 1.7 & 17.9 & 0.68 & 0.75 & 1.1 + & core m & 06 14 37.0 & 13 49 36 & & 1.7 & 17.6 & 0.72 & 3.33 & 1.3 + & & & & & 1.5 & 18.7 & 0.77 & & 1.2 + & & & & & 2.2 & 20.2 & 1.06 & & 2.5 + & core n & 06 14 37.0 & 13 49 57 & & 2.4 & 19.4 & 0.55 & 0.63 & 1.4 + & core o & 06 14 37.1 & 13 50 00 & & 1.9 & 17.4 & 0.53 & 0.51 & 1.0 + & core p & 06 14 37.2 & 13 49 49 & & 1.9 & 19.1 & 1.11 & 1.14 & 2.2 + & core q & 06 14 37.3 & 13 49 57 & & 1.4 & 18.5 & 0.89 & 0.97 & 1.2 + & core r & 06 14 37.3 & 13 50 02 & & 2.5 & 21.6 & 0.51 & 0.58 & 1.4 + & core s & 06 14 37.4 & 13 49 50 & & 1.1 & 16.9 & 0.61 & 0.56 & 0.7 + & core t & 06 14 37.4 & 13 49 51 & & 1.6 & 21.3 & 0.46 & 0.53 & 0.7 + & core u & 06 14 37.6 & 13 49 32 & & 1.4 & 19.3 & 0.51 & 0.53 & 0.8 + & core v & 06 14 38.1 & 13 49 26 & & 1.6 & 17.8 & 0.51 & 1.19 & 0.9 + & & & & & 1.5 & 18.5 & 0.60 & & 1.0 + & core w & 06 14 38.8 & 13 49 28 & & 2.3 & 21.7 & 0.48 & 0.49 & 1.1 + & core x & 06 14 38.8 & 13 49 39 & & 0.8 & 19.2 & 1.62 & 1.71 & 1.3 + & core y & 06 14 39.0 & 13 49 28 & & 1.7 & 17.1 & 0.60 & 0.58 & 1.0 + & core z & 06 14 39.1 & 13 49 27 & & 2.5 & 18.0 & 0.43 & 0.43 & 1.1 + & core aa & 06 14 39.2 & 13 49 25 & & 3.0 & 20.4 & 0.60 & 0.63 & 1.9 + & core ab & 06 14 39.2 & 13 49 26 & & 2.5 & 16.7 & 0.63 & 0.70 & 1.6 + & & & & & & & & & + iras19442 + 2427 & core a & 19 46 18.2 & 24 35 34 & & 2.4 & 20.4 & 0.67 & 0.75 & 1.6 + & core b & 19 46 18.2 & 24 35 41 & & 1.6 & 27.1 & 0.53 & 0.51 & 0.9 + & core c & 19 46 18.3 & 24 35 41 & & 2.4 & 20.9 & 0.73 & 0.77 & 1.8 + & core d & 19 46 18.4 & 24 35 30 & & 3.6 & 21.0 & 1.21 & 1.31 & 4.4 + & core e & 19 46 18.4 & 24 35 03 & & 3.3 & 21.5 & 0.72 & 0.82 & 2.3 + & core f & 19 46 18.5 & 24 35 09 & & 6.1 & 20.3 & 0.72 & 0.80 & 4.3 + & core g & 19 46 18.6 & 24 35 09 & & 2.6 & 21.5 & 0.61 & 0.68 & 1.6 + & core h & 19 46 18.7 & 24 35 26 & & 2.3 & 19.6 & 0.77 & 0.75 & 1.8 + & core i & 19 46 18.8 & 24 35 25 & & 4.5 & 21.1 & 0.89 & 0.96 & 4.0 + & core j & 19 46 18.8 & 24 35 21 & & 3.8 & 20.4 & 1.06 & 1.11 & 4.1 + & core k & 19 46 19.1 & 24 35 03 & & 4.4 & 20.6 & 0.84 & 0.96 & 3.7 + & core l & 19 46 19.2 & 24 35 44 & & 1.3 & 21.3 & 0.84 & 0.80 & 1.1 + & core m & 19 46 19.3 & 24 35 04 & & 1.6 & 21.5 & 0.72 & 1.07 & 1.2 + & & & & & 1.2 & 22.1 & 0.67 & & 0.8 + & core n & 19 46 19.3 & 24 35 19 & & 4.6 & 19.8 & 1.31 & 1.55 & 6.2 + & core o & 19 46 19.4 & 24 35 33 & & 3.7 & 20.8 & 0.84 & 0.97 & 3.1 + & core p & 19 46 19.5 & 24 35 07 & & 1.4 & 22.9 & 0.82 & 0.90 & 1.1 + & core q & 19 46 19.5 & 24 35 59 & & 2.9 & 21.3 & 0.89 & 0.94 & 2.5 + & core r & 19 46 19.6 & 24 35 17 & & 3.5 & 20.9 & 1.04 & 1.23 & 3.8 + & core s & 19 46 19.6 & 24 35 56 & & 2.5 & 27.1 & 0.44 & 0.46 & 1.1 + & core t & 19 46 19.7 & 24 35 04 & & 2.5 & 20.1 & 1.01 & 0.96 & 2.6 + & core u & 19 46 19.7 & 24 35 31 & & 1.7 & 21.5 & 1.06 & 2.17 & 1.8 + & & & & & 2.3 & 22.3 & 1.14 & & 2.7 + & core v & 19 46 19.8 & 24 35 43 & & 1.3 & 24.7 & 0.61 & 0.68 & 0.8 + & core w & 19 46 19.9 & 24 35 59 & & 4.3 & 20.8 & 0.94 & 0.97 & 4.1 + & core x & 19 46 20.0 & 24 35 56 & & 1.5 & 24.9 & 0.78 & 0.90 & 1.2 + & core y & 19 46 20.1 & 24 35 08 & & 1.6 & 22.7 & 0.65 & 0.67 & 1.1 + & core z & 19 46 20.2 & 24 35 51 & & 2.8 & 21.1 & 0.78 & 0.75 & 2.2 + & core aa & 19 46 20.2 & 24 35 53 & & 3.2 & 23.8 & 1.64 & 1.74 & 5.4 + & core ab & 19 46 20.3 & 24 35 18 & & 2.1 & 22.9 & 0.97 & 1.02 & 2.0 + & core ac & 19 46 20.4 & 24 35 34 & & 6.8 & 23.7 & 1.81 & 1.94 & 12.2 + & core ad & 19 46 20.4 & 24 35 46 & & 2.3 & 24.7 & 0.70 & 0.78 & 1.6 + & core ae & 19 46 20.4 & 24 35 53 & & 2.4 & 21.5 & 0.73 & 0.77 & 1.8 + & core af & 19 46 20.5 & 24 35 19 & & 1.7 & 22.1 & 0.99 & 1.07 & 1.6 + & core ag & 19 46 20.6 & 24 35 06 & & 3.3 & 23.6 & 1.13 & 1.21 & 3.8 + & core ah & 19 46 20.7 & 24 35 41 & & 2.3 & 21.2 & 1.14 & 1.40 & 2.8 + & core ai & 19 46 20.8 & 24 35 20 & & 1.2 & 24.3 & 0.65 & 0.67 & 0.8 + & core aj & 19 46 21.1 & 24 35 42 & & 1.4 & 23.2 & 0.58 & 0.65 & 0.8 + & core ak & 19 46 22.0 & 24 35 46 & & 2.3 & 20.7 & 0.77 & 0.84 & 1.8 + & core al & 19 46 22.1 & 24 35 28 & & 6.3 & 20.8 & 0.53 & 0.63 & 3.3 + & & & & & & & & & + iras19446 + 2505 & core a & 19 46 47.3 & 25 12 34 & & 1.8 & 22.4 & 0.60 & 0.60 & 1.1 + & core b & 19 46 47.7 & 25 13 04 & & 1.5 & 22.0 & 0.84 & 0.92 & 1.2 + & core c & 19 46 47.9 & 25 12 51 & & 1.7 & 22.0 & 0.63 & 0.73 & 1.1 + & core d & 19 46 47.9 & 25 13 17 & & 1.6 & 21.8 & 1.74 & 1.79 & 2.9 + & core e & 19 46 48.0 & 25 12 52 & & 4.2 & 21.0 & 1.35 & 1.40 & 5.9 + & core f & 19 46 48.1 & 25 12 46 & & 1.9 & 21.7 & 0.73 & 0.73 & 1.4 + & core g & 19 46 48.1 & 25 12 47 & & 1.2 & 18.8 & 0.77 & 0.75 & 1.0 + & core h & 19 46 48.3 & 25 12 33 & & 1.3 & 19.8 & 0.67 & 0.75 & 0.9 + & core i & 19 46 48.5 & 25 12 32 & & 1.4 & 18.1 & 0.61 & 0.73 & 0.8 + & core j & 19 46 48.5 & 25 13 05 & & 1.6 & 19.9 & 0.67 & 0.70 & 1.1 + & core k & 19 46 48.6 & 25 12 37 & & 2.4 & 21.2 & 0.87 & 0.96 & 2.1 + & core l & 19 46 48.6 & 25 12 53 & & 2.6 & 22.2 & 1.13 & 1.26 & 3.0 + & core m & 19 46 48.7 & 25 12 36 & & 1.4 & 22.0 & 0.73 & 0.82 & 1.0 + & core n & 19 46 48.7 & 25 12 50 & & 1.4 & 18.8 & 0.82 & 0.90 & 1.1 + & core o & 19 46 48.8 & 25 12 30 & & 1.7 & 18.8 & 0.78 & 0.75 & 1.3 + & core p & 19 46 48.9 & 25 12 34 & & 1.5 & 18.4 & 0.92 & 1.04 & 1.5 + & core q & 19 46 48.9 & 25 12 49 & & 4.8 & 20.9 & 2.29 & 2.34 & 11.4 + & core r & 19 46 49.0 & 25 12 41 & & 2.5 & 21.3 & 1.09 & 1.19 & 2.8 + & core s & 19 46 49.1 & 25 12 34 & & 1.7 & 17.6 & 0.56 & 0.60 & 1.0 + & core t & 19 46 49.1 & 25 12 31 & & 3.3 & 20.6 & 1.54 & 1.57 & 5.2 + & core u & 19 46 49.1 & 25 13 04 & & 6.2 & 21.0 & 1.23 & 1.28 & 7.9 + & core v & 19 46 49.2 & 25 12 34 & & 2.6 & 21.8 & 1.25 & 1.31 & 3.4 + & core w & 19 46 49.3 & 25 12 50 & & 4.9 & 20.5 & 1.86 & 1.96 & 9.9 + & core x & 19 46 49.4 & 25 12 56 & & 3.3 & 18.6 & 2.15 & 2.27 & 7.4 + & core y & 19 46 49.4 & 25 13 06 & & 2.6 & 20.0 & 1.13 & 1.07 & 3.1 + & core z & 19 46 49.5 & 25 12 32 & & 2.1 & 19.5 & 0.73 & 0.77 & 1.7 + & core aa & 19 46 49.5 & 25 12 33 & & 1.7 & 17.6 & 0.68 & 0.78 & 1.3 + & core ab & 19 46 49.5 & 25 12 47 & & 1.7 & 22.0 & 0.70 & 0.77 & 1.2 + & core ac & 19 46 49.7 & 25 13 11 & & 1.4 & 19.4 & 1.07 & 1.14 & 1.5 + & core ad & 19 46 49.8 & 25 12 39 & & 6.8 & 21.3 & 1.83 & 1.96 & 12.8 + & core ae & 19 46 49.9 & 25 12 31 & & 2.0 & 17.6 & 0.94 & 1.07 & 2.0 + & core af & 19 46 49.9 & 25 12 34 & & 1.7 & 18.2 & 0.77 & 0.80 & 1.3 + & core ag & 19 46 49.9 & 25 13 14 & & 1.7 & 18.9 & 1.28 & 1.35 & 2.2 + & core ah & 19 46 50.0 & 25 12 28 & & 4.9 & 21.8 & 0.97 & 1.11 & 5.0 + & core ai & 19 46 50.0 & 25 13 02 & & 2.7 & 20.8 & 0.99 & 1.09 & 2.8 + & core aj & 19 46 50.1 & 25 12 38 & & 3.6 & 19.4 & 1.35 & 1.45 & 4.9 + & core ak & 19 46 50.2 & 25 12 40 & & 1.4 & 18.2 & 0.99 & 0.97 & 1.4 + & core al & 19 46 50.3 & 25 12 43 & & 1.3 & 16.7 & 0.97 & 1.01 & 1.2 + & core am & 19 46 50.5 & 25 12 28 & & 3.9 & 20.8 & 1.42 & 1.47 & 5.7 + & core an & 19 46 50.5 & 25 12 56 & & 1.7 & 19.1 & 1.70 & 1.71 & 2.8 + & core ao & 19 46 50.6 & 25 12 35 & & 3.2 & 19.4 & 1.33 & 1.38 & 4.4 + cccccccccc & & & & & & + region & 98ghz & 110ghz & name & @xmath106(j2000 ) & @xmath192(j2000 ) & & & & f.f.@xmath210 + & ( mjy ) & ( mjy ) & & ( h m s ) & ( @xmath193 @xmath194 @xmath195 ) & & & & ( mjy ) + iras05375 + 3540 & [email protected] & [email protected] & mcs a & 05 40 53.4 & 35 41 53 & & [email protected] & [email protected] & no + & & & mcs b & 05 40 53.7 & 35 41 40 & & [email protected] & [email protected] & 0.4 + & & & s235 a & 05 40 52.9 & 35 42 16 & & [email protected] & [email protected] & 250 + & & & & & & & & & + iras06056 + 2131 & [email protected] & [email protected] & mcs a & 06 08 40.3 & 21 31 01 & & [email protected] & [email protected] & no + & & & mcs b & 06 08 40.6 & 21 31 07 & & [email protected] & [email protected] & 0.7 + & & & & & & & & & + iras06061 + 2151 & [email protected] & [email protected] & mcs a & 06 09 07.0 & 21 50 33 & & [email protected] & [email protected] & 3.4 + & & & mcs b & 06 09 07.0 & 21 50 42 & & [email protected] & [email protected] & 0.6 + & & & & & & & & & + iras06117 + 1350 & [email protected] & [email protected] & mcs a & 06 14 37.0 & 13 49 36 & & [email protected] & [email protected] & no ( ? ) + & & & s269 & 06 14 37.9 & 13 49 38 & & [email protected] & @xmath212 8.7 & 640 + & & & & & & & & & + iras19442 + 2427 & [email protected] & [email protected] & mcs a & 19 46 19.4 & 24 35 24 & & [email protected] & [email protected] & no + & & & mcs b & 19 46 20.6 & 24 35 39 & & [email protected] & [email protected] & no + & & & mcs c & 19 46 21.3 & 24 35 47 & & [email protected] & [email protected] & no + & & & s87 all & 19 46 20.1 & 24 35 29 & & [email protected] & [email protected] & 570 + & & & s87 core & ... & ... & & [email protected] & [email protected] & 29.7 + & & & & & & & & & + iras19446 + 2505 & 4200@xmath21130 & 3400@xmath21130 & mcs a & 19 46 46.6 & 25 13 01 & & [email protected] & 151@xmath21120 & no + & & & mcs b & 19 46 47.5 & 25 13 14 & & 219@xmath21117 & 157@xmath21127 & no + & & & mcs c & 19 46 48.4 & 25 12 20 & & 867@xmath21129 & 934@xmath21155 & no + & & & s88b1 & 19 46 48.2 & 25 12 42 & & 2180@xmath21122 & 1490@xmath21137 & 2500 + & & & s88b2 & 19 46 49.1 & 25 12 47 & & 574@xmath21110 & 394@xmath21116 & 640 + cclcccccccc iras05375 + 3540 & 44 & core a & 2.3 & 0.035 & 2.6 & 6.2 & 2.1 & ... & 1 & ... + & & core b & 3.6 & 0.027 & 2.9 & 5.2 & 5.2 & ... & ... & ... + & & core c & 3.2 & 0.034 & 3.7 & 9.9 & 3.2 & ... & ... & ... + & & core d & 3.1 & 0.020 & 1.8 & 3.0 & 7.8 & ... & ... & ... + & & core e & 1.1 & 0.026 & 0.9 & 1.7 & 1.7 & ... & ... & ... + & & core f & 0.9 & 0.019 & 0.5 & 1.1 & 2.3 & ... & ... & ... + & & core g & 1.8 & 0.022 & 1.2 & 1.6 & 3.8 & ... & 1 & ... + & & core h & 3.7 & 0.027 & 3.2 & 6.9 & 5.7 & b & 1 & 1 + & & core i & 5.1 & 0.042 & 6.2 & 16.8 & 2.9 & ... & ... & ... + & & core j & 8.9 & 0.026 & 6.5 & 9.4 & 12.8 & a & 1 & ... + & & core k & 2.6 & 0.022 & 1.8 & 2.7 & 5.8 & ... & ... & ... + & & core l & 2.0 & 0.015 & 1.2 & 1.8 & 12.1 & ... & ... & ... + & & core m & 2.0 & 0.045 & 4.6 & 5.3 & 1.8 & ... & ... & ... + & & core n & 3.1 & 0.047 & 6.5 & 7.4 & 2.2 & ... & ... & ... + & & core o & 1.0 & 0.010 & 0.5 & ... @xmath213 & ... @xmath213 & ... & ... & ... + & & core p & 5.0 & 0.035 & 6.1 & 6.4 & 5.0 & ... & ... & ... + & & core q & 1.7 & 0.022 & 1.3 & 2.5 & 4.3 & ... & ... & ... + & & core r & 2.4 & 0.026 & 1.7 & 8.0 & 3.4 & ... & ... & ... + & & core s & 3.4 & 0.043 & 5.2 & 4.6 & 2.3 & ... & 1 & ... + & & core t & 6.1 & 0.035 & 6.4 & 4.3 & 5.2 & ... & 1 & ... + & & core u & 1.9 & 0.021 & 1.3 & 1.2 & 4.9 & ... & ... & ... + & & & & & & & & & & + iras06056 + 2131 & 35 & core a & 5.0 & 0.061 & 16.7 & 12.1 & 2.6 & ... & ... & ... + & & core b & 1.2 & 0.035 & 2.5 & 9.8 & 2.0 & ... & ... & ... + & & core c & 4.6 & 0.042 & 15.5 & 23.1 & 7.3 & ... & ... & ... + & & core d & 3.5 & 0.049 & 9.6 & 15.9 & 2.8 & ... & ... & ... + & & core e & 10.9 & 0.071 & 54.1 & 133.8 & 5.3 & a & 1 & ... + & & core f & 1.7 & 0.060 & 6.5 & 11.0 & 1.1 & ... & 1 & ... + & & core g & 5.0 & 0.050 & 17.0 & 29.2 & 4.7 & ... & ... & ... + & & core h & 1.1 & 0.032 & 2.4 & 3.4 & 2.6 & ... & ... & ... + & & core i & 1.4 & 0.026 & 2.5 & 3.8 & 4.9 & ... & ... & ... + & & core j & 6.6 & 0.076 & 35.6 & 130.5 & 2.8 & b & ... & 2 + & & core k & 3.0 & 0.066 & 15.8 & 34.0 & 1.9 & ... & 1 & ... + & & core l & 1.0 & 0.043 & 2.9 & 4.2 & 1.3 & ... & ... & ... + & & core m & 1.1 & 0.068 & 6.3 & 8.8 & 0.6 & ... & 1 & ... + & & core n & 1.4 & 0.021 & 3.8 & ... @xmath213 & ... @xmath213 & ... & ... & ... + & & core o & 4.3 & 0.049 & 12.8 & 17.7 & 3.8 & ... & ... & ... + & & core p & 1.8 & 0.028 & 3.5 & 12.0 & 5.5 & ... & ... & ... + & & core q & 1.8 & 0.035 & 3.8 & 3.8 & 3.1 & ... & ... & ... + & & core r & 2.6 & 0.032 & 5.0 & 5.1 & 5.4 & ... & ... & ... + & & core s & 3.4 & 0.064 & 14.4 & 17.5 & 1.9 & ... & ... & ... + & & & & & & & & & & + iras06061 + 2151 & 39 & core a & 2.1 & 0.036 & 4.2 & 9.3 & 3.1 & ... & ... & ... + & & core b & 1.9 & 0.036 & 4.4 & 6.4 & 3.3 & ... & 1 & ... + & & core c & 0.9 & 0.040 & 2.3 & 1.8 & 1.2 & ... & ... & ... + & & core d & 1.5 & 0.043 & 3.9 & 11.4 & 1.7 & ... & ... & ... + & & core e & 1.8 & 0.045 & 5.4 & 6.8 & 2.1 & ... & 1 & ... + & & core f & 1.6 & 0.050 & 4.8 & 7.9 & 1.3 & ... & ... & ... + & & core g & 1.4 & 0.037 & 3.4 & 6.3 & 2.3 & ... & ... & ... + & & core h & 2.4 & 0.021 & 5.0 & ... @xmath213 & ... @xmath213 & ... & ... & ... + & & core i & 1.8 & 0.057 & 6.7 & 7.3 & 1.3 & ... & ... & ... + & & core j & 1.2 & 0.051 & 4.0 & 3.6 & 1.0 & ... & ... & ... + & & core k & 1.1 & 0.039 & 2.9 & 2.9 & 1.7 & ... & ... & ... + & & core l & 2.0 & 0.026 & 3.9 & 4.0 & 7.7 & ... & ... & ... + & & core m & 1.5 & 0.036 & 4.1 & 11.4 & 3.0 & ... & ... & ... + & & core n & 1.2 & 0.027 & 2.8 & 2.1 & 4.9 & ... & 1 & ... + & & core o & 7.5 & 0.043 & 19.4 & 94.4 & 8.5 & b & 2 & 2 + & & core p & 2.6 & 0.040 & 5.4 & 62.3 & 3.0 & a & 1 & 1 + & & core q & 1.8 & 0.059 & 6.2 & 11.7 & 1.0 & ... & ... & ... + & & core r & 1.3 & 0.030 & 3.1 & 3.7 & 4.0 & ... & 1 & ... + & & core s & 3.2 & 0.051 & 9.8 & 17.9 & 2.6 & ... & ... & ... + & & core t & 3.8 & 0.035 & 8.3 & 12.6 & 6.7 & ... & ... & ... + & & core u & 3.1 & 0.067 & 12.3 & 9.0 & 1.4 & ... & ... & ... + & & core v & 1.6 & 0.035 & 4.0 & 9.8 & 3.2 & ... & ... & ... + & & core w & 3.4 & 0.051 & 10.3 & 15.6 & 2.7 & ... & ... & ... + & & core x & 3.3 & 0.090 & 19.4 & 32.5 & 0.9 & ... & ... & ... + & & & & & & & & & & + iras06117 + 1350 & 39 & core a & 1.7 & 0.024 & 3.6 & 3.0 & 9.1 & ... & ... & ... + & & core b & 2.2 & 0.038 & 6.1 & 8.1 & 3.9 & ... & ... & ... + & & core c & 1.2 & 0.037 & 3.6 & 3.4 & 2.5 & ... & ... & ... + & & core d & 1.1 & 0.023 & 2.5 & 2.4 & 7.2 & ... & ... & ... + & & core e & 1.6 & 0.018 & 2.7 & ... @xmath213 & ... @xmath213 & ... & ... & ... + & & core f & 1.4 & 0.039 & 4.0 & 1.7 & 2.3 & ... & ... & ... + & & core g & 1.2 & 0.015 & 2.3 & ... @xmath213 & ... @xmath213 & ... & ... & ... + & & core h & 0.8 & 0.025 & 1.7 & 2.9 & 3.9 & ... & ... & ... + & & core i & 1.2 & 0.021 & 2.3 & ... @xmath213 & ... @xmath213 & ... & ... & ... + & & core j & 1.8 & 0.059 & 7.1 & 29.1 & 1.2 & ... & ... & ... + & & core k & 1.5 & 0.039 & 4.3 & 5.0 & 2.5 & ... & ... & ... + & & core l & 1.4 & 0.059 & 5.6 & 7.0 & 1.0 & ... & ... & ... + & & core m & 5.9 & 0.062 & 24.3 & 143.4 & 3.5 & a & 1 & ... + & & core n & 1.7 & 0.059 & 6.5 & 4.9 & 1.1 & ... & ... & ... + & & core o & 1.3 & 0.039 & 3.2 & 2.1 & 1.9 & ... & ... & ... + & & core p & 2.7 & 0.066 & 10.9 & 18.0 & 1.3 & ... & ... & ... + & & core q & 1.6 & 0.054 & 6.4 & 10.7 & 1.4 & ... & ... & ... + & & core r & 1.7 & 0.043 & 4.6 & 3.0 & 2.0 & ... & ... & ... + & & core s & 0.8 & 0.014 & 1.6 & ... @xmath213 & ... @xmath213 & ... & ... & ... + & & core t & 0.9 & 0.024 & 3.2 & 1.4 & 8.0 & ... & ... & ... + & & core u & 0.9 & 0.030 & 2.1 & 1.8 & 2.7 & ... & ... & ... + & & core v & 2.1 & 0.068 & 10.1 & 20.3 & 1.1 & ... & ... & ... + & & core w & 1.4 & 0.015 & 2.7 & ... @xmath213 & ... @xmath213 & ... & ... & ... + & & core x & 1.6 & 0.042 & 4.2 & 25.6 & 2.0 & ... & ... & ... + & & core y & 1.3 & 0.017 & 2.5 & ... @xmath213 & ... @xmath213 & ... & ... & ... + & & core z & 1.3 & 0.044 & 3.8 & 1.7 & 1.5 & ... & ... & ... + & & core aa & 2.3 & 0.022 & 4.8 & ... @xmath213 & ... @xmath213 & ... & ... & ... + & & core ab & 2.0 & 0.020 & 3.8 & ... @xmath213 & ... @xmath213 & ... & ... & ... + & & & & & & & & & & + iras19442 + 2427 & 44 & core a & 2.2 & 0.027 & 2.4 & 3.2 & 4.2 & ... & ... & ... + & & core b & 1.2 & 0.031 & 1.1 & 1.7 & 1.3 & ... & ... & ... + & & core c & 2.5 & 0.028 & 2.8 & 3.5 & 4.5 & ... & ... & ... + & & core d & 6.1 & 0.039 & 9.1 & 14.1 & 5.3 & ... & ... & ... + & & core e & 3.3 & 0.047 & 6.2 & 6.6 & 2.1 & ... & ... & ... + & & core f & 6.4 & 0.051 & 12.4 & 6.9 & 3.3 & ... & ... & ... + & & core g & 2.2 & 0.025 & 2.1 & 2.4 & 4.7 & ... & ... & ... + & & core h & 2.5 & 0.026 & 1.8 & 3.1 & 3.6 & ... & ... & ... + & & core i & 5.7 & 0.038 & 4.9 & 7.3 & 3.1 & ... & ... & ... + & & core j & 5.7 & 0.050 & 9.8 & 12.9 & 2.7 & ... & ... & ... + & & core k & 5.3 & 0.056 & 10.4 & 10.7 & 2.1 & ... & ... & ... + & & core l & 1.5 & 0.012 & 0.8 & ... @xmath213 & ... @xmath213 & ... & ... & ... + & & core m & 2.7 & 0.032 & 3.6 & 7.7 & 3.8 & ... & 1 & ... + & & core n & 8.8 & 0.051 & 15.7 & 25.7 & 4.1 & a & ( 1)@xmath214 & ... + & & core o & 4.4 & 0.028 & 4.2 & 5.5 & 6.7 & ... & ... & ... + & & core p & 1.5 & 0.027 & 1.4 & 4.6 & 2.5 & ... & ... & ... + & & core q & 3.6 & 0.036 & 5.3 & 6.6 & 3.9 & ... & ... & ... + & & core r & 5.2 & 0.049 & 10.4 & 15.5 & 3.1 & a & ( 1)@xmath214 & ... + & & core s & 1.6 & 0.015 & 0.9 & 0.7 & 9.6 & ... & ... & ... + & & core t & 3.5 & 0.034 & 4.2 & 6.5 & 3.7 & ... & ... & ... + & & core u & 6.1 & 0.056 & 14.4 & 54.9 & 2.9 & ... & 1 & ... + & & core v & 1.1 & 0.030 & 1.3 & 2.9 & 1.7 & ... & ... & ... + & & core w & 5.8 & 0.043 & 8.0 & 8.5 & 3.5 & ... & 1 & ... + & & core x & 1.6 & 0.033 & 2.4 & 5.6 & 2.3 & ... & ... & ... + & & core y & 1.5 & 0.013 & 0.9 & ... @xmath213 & ... @xmath213 & ... & ... & ... + & & core z & 3.1 & 0.036 & 4.1 & 4.2 & 3.0 & ... & ... & ... + & & core aa & 7.5 & 0.048 & 9.7 & 30.4 & 3.1 & ... & ... & ... + & & core ab & 2.8 & 0.051 & 4.9 & 11.2 & 1.3 & ... & ... & ... + & & core ac & 18.2 & 0.047 & 14.1 & 37.2 & 4.7 & ... & ... & ... + & & core ad & 2.3 & 0.029 & 2.3 & 3.7 & 3.3 & ... & ... & ... + & & core ae & 2.4 & 0.025 & 2.2 & 3.1 & 5.0 & ... & ... & ... + & & core af & 2.3 & 0.014 & 1.5 & 3.4 & 18.6 & ... & ... & ... + & & core ag & 5.2 & 0.046 & 7.6 & 14.1 & 2.7 & ... & ... & ... + & & core ah & 3.7 & 0.032 & 4.2 & 13.9 & 4.4 & b & ... & ... + & & core ai & 1.0 & 0.020 & 1.1 & 1.9 & 4.8 & ... & ... & ... + & & core aj & 1.1 & 0.019 & 0.7 & 1.7 & 3.8 & ... & ... & ... + & & core ak & 2.4 & 0.022 & 1.9 & 3.2 & 6.1 & ... & 1 & ... + & & core al & 5.0 & 0.039 & 5.8 & 3.3 & 3.4 & ... & 1 & ... + & & & & & & & & & & + iras19446 + 2505 & 60 & core a & 2.0 & 0.010 & 1.0 & ... @xmath213 & ... @xmath213 & ... & ... & ... + & & core b & 2.2 & 0.025 & 1.9 & 4.4 & 4.3 & ... & ... & ... + & & core c & 2.0 & 0.018 & 1.7 & 2.0 & 10.0 & ... & ... & ... + & & core d & 5.2 & 0.036 & 8.8 & 24.1 & 6.0 & ... & 1 & ... + & & core e & 10.6 & 0.042 & 15.0 & 17.2 & 7.1 & ... & ... & ... + & & core f & 2.5 & 0.010 & 1.4 & ... @xmath213 & ... @xmath213 & ... & ... & ... + & & core g & 1.7 & 0.019 & 1.3 & 2.2 & 6.5 & ... & ... & ... + & & core h & 1.6 & 0.022 & 1.5 & 2.6 & 4.8 & ... & ... & ... + & & core i & 1.6 & 0.017 & 1.1 & 1.9 & 7.9 & ... & ... & ... + & & core j & 1.9 & 0.023 & 1.4 & 2.4 & 3.9 & ... & ... & ... + & & core k & 3.8 & 0.026 & 3.4 & 5.0 & 6.7 & ... & ... & ... + & & core l & 5.4 & 0.040 & 7.7 & 13.3 & 4.2 & ... & ... & ... + & & core m & 1.8 & 0.024 & 1.7 & 3.4 & 4.3 & ... & ... & ... + & & core n & 2.0 & 0.015 & 1.3 & 2.6 & 13.2 & ... & ... & ... + & & core o & 2.4 & 0.027 & 1.9 & 3.2 & 3.3 & ... & ... & ... + & & core p & 2.5 & 0.026 & 2.4 & 5.9 & 4.8 & ... & ... & ... + & & core q & 20.6 & 0.032 & 19.7 & 36.5 & 20.9 & ... & ... & ... + & & core r & 5.0 & 0.028 & 5.2 & 8.4 & 8.2 & ... & ... & ... + & & core s & 1.8 & 0.015 & 1.1 & 1.1 & 10.8 & ... & ... & ... + & & core t & 9.4 & 0.033 & 10.0 & 17.0 & 9.6 & ... & ... & ... + & & core u & 14.4 & 0.027 & 11.5 & 9.2 & 20.3 & ... & ... & ... + & & core v & 6.0 & 0.023 & 4.6 & 8.3 & 13.3 & ... & ... & ... + & & core w & 17.1 & 0.042 & 25.8 & 33.8 & 12.1 & ... & ... & ... + & & core x & 13.2 & 0.043 & 19.3 & 46.3 & 8.5 & ... & ... & ... + & & core y & 5.4 & 0.038 & 6.7 & 9.2 & 4.2 & ... & ... & ... + & & core z & 2.9 & 0.015 & 1.8 & 1.9 & 18.2 & ... & ... & ... + & & core aa & 2.1 & 0.010 & 1.2 & ... @xmath213 & ... @xmath213 & ... & 1 & ... + & & core ab & 2.1 & 0.011 & 1.1 & ... @xmath213 & ... @xmath213 & ... & ... & ... + & & core ac & 2.8 & 0.023 & 2.2 & 6.3 & 6.1 & ... & ... & ... + & & core ad & 23.6 & 0.046 & 38.4 & 37.0 & 13.7 & ... & ... & ... + & & core ae & 3.4 & 0.019 & 2.3 & 4.6 & 11.6 & ... & ... & ... + & & core af & 2.4 & 0.030 & 2.1 & 4.0 & 2.7 & ... & ... & ... + & & core ag & 3.9 & 0.021 & 2.7 & 8.0 & 10.0 & ... & ... & ... + & & core ah & 8.9 & 0.028 & 8.1 & 7.2 & 12.8 & ... & ... & ... + & & core ai & 4.9 & 0.013 & 3.0 & ... @xmath213 & ... @xmath213 & ... & 1 & ... + & & core aj & 8.8 & 0.045 & 15.1 & 19.8 & 5.8 & ... & ... & ... + & & core ak & 2.4 & 0.018 & 1.6 & 3.6 & 9.3 & ... & ... & ... + & & core al & 2.2 & 0.017 & 1.5 & 3.6 & 10.7 & ... & ... & ... + & & core am & 10.2 & 0.023 & 7.3 & 10.4 & 20.9 & ... & ... & ... + & & core an & 5.1 & 0.020 & 3.7 & 12.2 & 16.3 & ... & ... & ... + & & core ao & 7.8 & 0.024 & 6.2 & 9.6 & 15.6 & ... & ... & ... + cccccccccc region & mcs & index@xmath189 & type@xmath190 & @xmath75 & @xmath82 & @xmath78 & @xmath215 & + & & & & ( pc ) & ( k ) & ( @xmath3 ) & ( @xmath216 @xmath9 ) & nir@xmath217 & ccs@xmath218 + iras05375 + 3540 & mcs a & [email protected] & d.t . & 0.032 & 31.5 & 35.0 & 44.7 & 1 & ... + & mcs b & [email protected] & d.t . & 0.019 & 31.5 & 9.3 & 46.1 & 1 & 1 + & & & & & & & + iras06056 + 2131 & mcs a & [email protected] & d.t . & 0.055 & 56.0 & 111 & 22.9 & 1 & ... + & mcs b & [email protected] & d.t . & 0.037 & 54.0 & 64.1 & 42.6 & ... & 2 + & & & & & & & + iras06061 + 2151 & mcs a & [email protected] & d.t . & 0.021 & 60.0 & 2.7 & 10.1 & 1 & 1 + & mcs b & [email protected] & d.t . & 0.033 & 60.0 & 35.6 & 35.8 & 2 & 2 + & & & & & & & + iras06117 + 1350 & mcs a & [email protected] & d.t . & 0.048 & 66.0 & 43.9 & 13.5 & 1 & ... + & & & & & & & + iras19442 + 2427 & mcs a & [email protected] & d.t . & 0.021 & 43.5 & 9.1 & 34.1 & ( 1)@xmath213 & ... + & mcs b & [email protected] & d.t . & 0.018 & 77.0 & 14.1 & 80.6 & ... & ... + & mcs c & [email protected] & f.f . & 0.032 & & & & ... & ... + & & & & & & & + iras19446 + 2505 & mcs a & [email protected] & f.f . & 0.050 & & & & ... & ... + & mcs b & @xmath2190.5@xmath220 & f.f . & 0.057 & & & & ... & ... + & mcs c & [email protected] & f.f . & 0.110 & & & & ... & ... + cccccccc region & clump & @xmath126 & @xmath221 & @xmath175 & @xmath60 & @xmath161 & core@xmath189 + & & & ( pc ) & ( @xmath3 ) & ( @xmath3 ) & ( @xmath222 @xmath9 ) + iras05375 + 3540 & clump b & 1.5 & 0.16 & 80 & 79 & 6.8 & a , b + & clump c & 2.1 & 0.21 & 230 & 200 & 8.6 & c - m , o + & clump d & 1.6 & 0.21 & 170 & 120 & 6.4 & n , p - u + & & & & & & + iras06056 + 2131 & clump a & 2.6 & 0.41 & 1000 & 560 & 5.0 & a - s + & & & & & & + iras06061 + 2151 & clump b & 2.9 & 0.61 & 1300 & 1100 & 2.0 & a - w + & & & & & & + iras06117 + 1350 & clump c & 2.3 & 0.37 & 380 & 390 & 2.6 & d - f , k , l + & clump d & 2.6 & 0.18 & 90 & 250 & 5.4 & g - j , m , u + & & & & & & + iras19442 + 2427 & clump a & 2.6 & 0.29 & 380 & 400 & 5.6 & a , c - o , r , t , u , af , + & clump b & 2.2 & 0.26 & 260 & 270 & 5.1 & b , p , s , v , x , y , aa - ad , ag , ai , aj + & clump c & 1.6 & 0.34 & 290 & 180 & 2.6 & q , w , z , ae , ah , ak , al + & & & & & & + iras19446 + 2505 & clump b & 2.6 & 0.22 & 460 & 320 & 15.0 & a - ag , ai - al , an + & clump c & 2.5 & 0.18 & 210 & 230 & 12.5 & ah , aj , ak , ao +	we present the results of c@xmath0o observations by the nobeyama millimeter array toward dense clumps with radii of @xmath1 pc in six cluster - forming regions including massive ( proto)stars . we identified 171 cores , whose radius , line width , and molecular mass range from 0.01 to 0.09 pc , 0.43 to 3.33 km s@xmath2 , and 0.5 to 54.1 @xmath3 , respectively . many cores with various line widths exist in one clump , and the index of the line width radius relationship of the cores and the parental clump differs from core to core in the clump . this indicates that the degree of dissipation of the turbulent motion varies for each core in one clump . although the mass of the cores increases with the line width , most cores are gravitationally bound by the external pressure . in addition , the line width and the external pressure of the cores tend to decrease with the distance from the center of the clump , and these dependencies may be caused by the inner h@xmath4 density structure of the clump that affects the physical properties of the cores . moreover , the number density of the cores and the number density of young ( proto)stars have a similar relationship to the average h@xmath4 density of the clumps . thus , our findings suggest that the cluster is formed in the clump through the formation of such multiple cores , whose physical properties would have been strongly related to the h@xmath4 density structure and the turbulent motion of the clump .
observations have shown that there exists a supermassive black hole , sgr a * , located at the galactic center ( gc ) . the mass of the black hole is about @xmath2 ( @xcite ; @xcite ; @xcite ) . because of its proximity , sgr a * is regarded as the best laboratory of studying black hole accretion . numerous observations have been conducted and abundant data has been obtained ( see recent reviews by @xcite ; falcke & markoff 2013 ; yuan & narayan 2014 ) . the source is quite dim currently , with a bolometric luminosity of only about @xmath3 . the mass accretion rate at the bondi radius has been estimated by combining the _ chandra _ observation and the bondi accretion theory , which is @xmath4 ( @xcite ) . the bolometric luminosity would be 5 orders of magnitude higher if the accretion were in the mode of the standard thin disk . amount of theoretical studies in the past 20 years have revealed that the advection - dominated accretion flow ( adaf ) can explain this puzzle ( @xcite ) . specifically , the low - luminosity of sgr a * is because of two reasons . one is the intrinsic low radiative efficiency of adaf because of energy advection ( @xcite ; @xcite ) . another important reason is because of the existence of strong wind ( or outflow ) , i.e. , @xmath5 of the matter captured at the bondi radius are lost ( @xcite ; @xcite ; @xcite ) . the existence of wind has been confirmed by the radio polarization observations ( e.g. , @xcite ; @xcite ; @xcite ) , and more recently by the _ chandra _ observation to the emission lines from the accretion flow in sgr a * ( @xcite ) . yuan & narayan ( 2014 ) presented the most recent review on the hot accretion flow and its various astrophysical applications , including on sgr a. one particularly interesting thing is that many observational evidences show that the activity of sgr a was very likely much stronger in the past than the current stage . these observations suggest that sgr a * has perhaps undergone multiple past epochs of enhanced activity on different timescales . here we only focus on relatively long timescales . these evidences were summarized in @xcite , and later discussed in other works ( e.g. , @xcite ; @xcite ; @xcite ) . these evidences include : 1 ) orders of magnitude higher x - ray luminosity ( compared to the present value ) required to explain the fluorescent x - ray emission reflected from cold iron atoms in the giant molecular cloud sgr b2 ( @xcite ; @xcite ; @xcite ) ; 2 ) a new x - ray reflection nebula associated with sgr c detected by _ asca _ ( @xcite ) ; 3 ) the ionized halo surrounding sgr a * ( maeda et al . 2002 ) ; 4 ) galactic center lobe ( gcl , @xcite ) ; 5 ) expanding molecular ring ( emr , @xcite ; @xcite ) ; 6 ) north polar spur ( nps , @xcite ; @xcite ) ; 7 ) the 8 kev diffuse x - ray emission in the center ( @xcite ) ; 8) the excess of h@xmath6 emission of magellanic stream ( @xcite ) ; 9 ) the _ suzaku _ observations to the nps ( @xcite ) . @xcite found that to explain the former seven observations mentioned above , the characteristic x - ray luminosity of sgr a * should be @xmath7 several hundred years ago , and such an activity should last for @xmath1 yr . for such a luminosity , the accretion should be well in the regime of hot accretion rather than the standard thin disk ( yuan & narayan 2014 ) . correspondingly , the mass accretion rate should be @xmath0 times higher than the present value ( @xcite ) . other possibilities of the past activity have also been proposed . for example , the bolometric luminosity in the past millions years estimated by @xcite based on the 8th evidence mentioned above is much higher , @xmath8 . the timescale of the activity is shorter , and it was active 1@xmath93 myr ago . yet another possibility is as follows . a star formation event has been observed and it is believed to occur at @xmath10 yr ago on scales of @xmath11 [email protected] pc from the smbh ( e.g. , @xcite ; @xcite ) . if the past activity of sgr a * occur concurrently with this event , this would imply a strong activity of sgr a * occurred @xmath12 myr ago ( @xcite ) . in summary , so far we still lack a consensus on the past activity of sgr a. yet perhaps another evidence for the past activity of sgr a is the _ fermi _ bubbles recently detected . using the _ fermi_-lat , @xcite discovered two giant gamma - ray bubbles located above and below the galactic plane ( also refer to @xcite for the recent observations ) . in galactic coordinates ( l , b ) , the height of each bubble is about @xmath13 , and the width is about @xmath14 . the surface brightness looks uniform , and the edge looks sharp . the total luminosity of the bubbles is @xmath15 in 1@xmath9100 gev band . the total energy of the two bubbles is estimated to be @xmath16 erg . many theoretical models have been proposed since the discovery of the _ fermi _ bubbles . in the `` hadronic '' model , the formation is explained as due to a population of relic cosmic ray protons injected by processes associated with extremely long time scale and high areal density star formation in the galactic center ( @xcite ) . in the `` leptonic '' scenario the @xmath17-ray emission comes from the inverse compton scattering between relativistic electrons ( also often called as cosmic ray ) and seed photons . the seed photons may be the cosmic microwave background , but the origin of relativistic electrons are different in different models . they can come from fermi-1st order acceleration on shock front formed in the periodic star capture processes by sgr a * ( @xcite ) , the fermi-2nd order acceleration through stochastic scattering by plasma instabilities ( @xcite ) , directly from the jet ( @xcite ; @xcite ; @xcite ; @xcite ) , or from outflows driven by the past star formation ( @xcite ) . among these models , there are two models which are physically most relevant to the model we propose in the present paper . they are the `` jet '' model ( @xcite ; @xcite ) and the `` quasar outflow '' model ( @xcite ; @xcite ) . in the former , it is suggested that the bubbles are created by agn jet which happened about 2 myr ago . after that , cosmic rays ( crs ) carried by jet diffuse to today s morphology . yang et al . ( 2012 , 2013 ) developed the jet model by including magnetic field . they showed that the suppression of the diffusion of crs along the direction across the edge is caused by the magnetic field configuration . this is because inside the bubbles , the magnetic field is mainly radial , but just outside of the bubble and close to the edge , the field is mainly in the parallel direction . one problem , as pointed out by @xcite , is that they must require the jet direction to be perpendicular to the plane of the galaxy , which seems to be unlikely , given the general absence of correlation between the direction of jets and galaxy planes and the observed direction of the stellar disk in the galaxy . in addition , the velocity required in the jet model is as low as @xmath18 and the mass loss rate in the jet is in general as high as super - eddington . another model is the `` quasar outflow '' model proposed in @xcite and @xcite . in this model , sgr a * is again assumed to be very active in the past , with mildly super - eddington accretion rate 6 myr ago and duration of the activity being 1 myr . under such a high luminosity , quasi - spherical outflow will be driven by the strong radiation pressure from this quasar ( @xcite ) , which can result in the formation of the _ fermi _ bubbles . in this model , the existence of the well - known central molecular zone ( cmz ) in the gc region plays an important role in collimating the outflow and forming the morphology of the bubbles . @xcite pointed out that the expansion velocity derived by the _ suzaku _ observation is lower than the advocated values by both the jet and quasar outflow models by a factor of 5 and 2 respectively . assuming that sgr a * was in an active state as suggested by totani ( 2006 ) , in this paper we investigate whether the _ fermi _ bubbles can be inflated by the wind launched from the hot accretion flow by performing numerical simulations . in 2 , we briefly introduce some background on the accretion flow and wind , and present an analytical solution for the interaction between the winds and ism to be used to understand our numerical simulation results . the numerical simulations approach and the results are presented in 3 and 4 , respectively . we then summarize in 5 . the accretion flow in sgr a * in the current stage is relatively simple , namely the whole accretion flow is hot , ranging from the bondi radius to the black hole horizon . however , if the accretion rate is @xmath0 higher , as estimated by totani ( 2006 ) , this simple picture needs to be modified . numerous observational and theoretical studies have shown that the accretion flow should consist of an outer thin disk and an inner hot accretion flow . the boundary between the truncated thin disk and the hot accretion flow is called the transition radius ( @xmath19 ) . some works have been done on the physical mechanism of the transition . although this question is still not completely solved , we now have a consensus that the value of @xmath19 should decrease with increasing mass accretion rate . this is supported by the modeling to some low - luminosity agns and the hard state of black hole x - ray binaries , which is summarized in @xcite . for hot accretion flow , the mass accretion rate is a function of radius because of the mass loss in the wind throughout the disk ( refer to [ wind ] ) . the current net mass accretion rate at the horizon of the black hole and at the bondi radius are @xmath20 and @xmath21 , respectively ( @xcite ) . here @xmath22 is defined as the eddington accretion rate . according to @xcite , the mass accretion rate close to the horizon of black hole in sgr a * should be @xmath23 during the past @xmath24 yr . for this value of accretion rate , given the theoretical uncertainty , @xmath25 would be a reasonable assumption , here @xmath26 is the schwarzschild radius of the black hole . note that there is some uncertainty in the value of @xmath19 . the mass accretion rate at @xmath25 is set to be @xmath27 in our favored model . this value is @xmath28 times higher than the present value , well within the range obtained in @xcite . as we have mentioned in 1 , one characteristic feature of hot accretion flow is that it is subject to strong wind . the existence of wind has been suggested in @xcite and later by @xcite . the hydrodynamic ( hd ) and magnetohydrodynamic ( mhd ) numerical simulation works by @xcite and @xcite showed that the mass inflow rate of the accretion flow decreases inward , which can be regarded as the pioneer works in the quantitative study of winds from hot accretion flow . this result is confirmed by many other subsequent works ( see review by @xcite ) . it was soon shown that the physical reason for the inward decrease of inflow rate is because of mass loss in wind which occurs in a wide range of radius throughout the accretion flow ( @xcite ; @xcite ; @xcite ; @xcite ) . the physical mechanism for the production of winds is found to be the combination of magnetocentrifugal force and the gradient of gas and magnetic pressure ( @xcite ; yuan et al . in preparation ) . while the existence of wind is evident , consensus on some quantitative features of the wind have not been reached . for example , @xcite argued that the mass flux of wind should be significant , comparable to the mass flux of inflow . this is much larger than the lower limit obtained in @xcite . in this work , we follow @xcite and assume that at @xmath19 , the mass flux of wind is roughly equal to the inflow rate there , i.e. , @xmath29 in most of our models except for runs `` g '' and `` h '' ( refer to table 1 ) . @xcite ( see also @xcite ) also estimated the terminal radial velocity of wind based on the conservation of the value of bernoulli parameter @xmath30 and found it is roughly half of the keplerian velocity at @xmath19 . however , that estimation should be regarded as the lower limit since magnetic field is not included in the analysis . our more recent study find that @xmath30 actually increases along the streamline when magnetic field is included ( yuan et al . 2014 , in preparation ) . in the present work we set the velocity of the wind to be @xmath31 our simulations indicate that there is some degeneracy between the mass flux and the velocity of winds . what really matters is the power of winds . the mass flux and velocity adopted above correspond to the power of wind @xmath32 . the next wind parameter is their angular distribution . in spherical coordinate , yuan et al . ( in preparation ) find that winds occupy a region @xmath33 and @xmath34 . given that the range is quite large , combining with the possibility that during the long timescale of @xmath24 yr the rotation axis of the accretion flow may have changed with time , in the present work , we simply assume that the winds are blown out isotropically . winds may also be launched from the truncated thin disk outside of @xmath19 . but the details of this process is poorly investigated at present . in this work , we assume that this part of wind is not important compared to the winds from the inner hot accretion flow . this is the main uncertainty of our model . the winds launched from the hot accretion flow are usually supersonic so they will interact with the interstellar medium and produce shocks . before we present the details of our simulation results , in this part we present some analytical solutions to this problem based on some simplifications , which is helpful to understanding our simulation results . here we assume a simple shock model formed by an isotropic wind punching into an isotropic distribution of interstellar medium ( ism ) . it is well - known that the region can be divided into the following four parts : 1 ) high speed wind ; 2 ) shocked wind ; 3 ) shocked ism ; 4 ) un - shocked ism gas . the interface between the shocked wind and shocked ism is called the contact discontinuity ( cd ) . in our case , since the cooling timescale of shocked winds is longer than the flow timescale , the shocked wind is a `` energy - driven '' flow rather than `` momentum - driven '' flow ( @xcite , @xcite , @xcite and @xcite ) . if we assume that the forward shock velocity @xmath35 is equal to the velocity of the shocked ism ( @xmath36 ) and the shocked ism region is so thin that @xmath37 , approximately we can obtain the following equations ( e.g. , @xcite , weaver et al . 1977 ) : @xmath38 e_{b}\approx \frac{4}{3}\pi r^{3}_{2 } \cdot \frac{3}{2 } p_{b } , \\ % [ 1em ] \frac{d}{dt}(\mc \rdot2)=4\pi r^{2}_2 p_{b } , \\ % [ 1em ] \mc=\int^{r_{2}}_{0 } \rho_{\texttt{ism } } 4\pi r^{2}dr , \\ % [ 1em ] p_{w}=\frac{1}{2}\mdotw v^{2}. \end{gathered}\ ] ] here , @xmath39 is the total energy of shocked wind , in which the internal energy is dominant , @xmath40 is the kinetic power of the un - shocked wind , @xmath41 is the gas pressure of shocked wind . in the case of weak shock , the shock velocity @xmath35 will be significantly higher than the velocity of the shocked ism , so our approximations may introduce large errors . assuming @xmath42 , in which _ a _ and _ n _ are both constants and @xmath43 , the solutions of the above equations are : @xmath44 p_{b}(t)=g(a , n)\cdot p_{w}^{\frac{2-n}{5-n } } t^{-\frac{n+4}{5-n } } , \label{equ_p } \\ % [ 1em ] f(a , n)=\left[\frac{(5-n)^{3}(3-n)}{14\pi a(7 - 2n)(11-n)}\right]^{\frac{1}{5-n } } , \label{equ_rfan } \\ g(a , n)=\frac{3a(7 - 2n)}{(5-n)^{2}(3-n)}\cdot f(a , n)^{2-n}. \end{gathered}\ ] ] here , @xmath45 is the radii of the forward shock , and roughly it can be used to represent the radii of cd . here we have neglected the gravity . this is because in our case the work done by overcoming the gravity is one order of magnitude lower than the injected energy from sgr a. our solution of @xmath45 is similar to the energy - driven solution in @xcite . besides , we find that when the density profile is assumed to be the same as @xcite , the velocity of the shocked ism is also close to their result . @ c c c c c c c c c c & @xmath46 & @xmath47 & @xmath48 is the velocity of the wind . ] & @xmath49 & @xmath50 & @xmath40 is the total kinetic power of the wind injected in @xmath51 of solid angle . ] & @xmath52 is the age of _ fermi _ bubbles . ] & @xmath53 is the duration of the quiescent state of sgr a in the final stage . ] + run & @xmath54 ) & ( @xmath55 ) & ( c ) & ( @xmath56 ) & ( @xmath57 ) & ( @xmath58 ) & ( myr ) & ( myr ) + a & @xmath59 & @xmath60 & 6.2% & 2.0 & 2.0% & 2.0 & 12.3 & - + b & @xmath59 & @xmath60 & 6.2% & 2.0 & 2.0% & 2.0 & 12.3 & 0.3 + c & @xmath59 & @xmath60 & 6.2% & 0 & 2.0% & 2.0 & 7.6 & - + d & @xmath59 & @xmath60 & 6.2% & 4.0 & 2.0% & 2.0 & 14.5 & - + e & @xmath59 & @xmath60 & 6.2% & 2.0 & 2.0% & 2.0 & 13.4 & - + f & @xmath61 & @xmath60 & 6.2% & 2.0 & 2.0% & 2.0 & 14.9 & - + g & @xmath59 & @xmath60 & 6.2% & 3.0 & 6.0% & 6.0 & 8.1 & - + h & @xmath59 & @xmath62 & 4.3% & 2.0 & 4.2% & 2.0 & 11.6 & - + [ table1 ] @ c c c c c c c c c c c & h / w & @xmath63 is the space - averaging temperature of _ fermi _ bubbles ] & @xmath64 is the space - averaging temperature of shocked ism . ] & @xmath65 is the internal energy of _ fermi _ bubbles . ] & @xmath66 is the kinetic energy of _ fermi _ bubbles . ] & @xmath67 is the internal energy of shocked ism . ] & @xmath68 is the kinetic energy of shocked ism . ] & @xmath69 is the total energy injected by sgr a * wind . ] & @xmath70 is the total mass injected from the origin . ] & @xmath71 is the total mass of _ fermi _ bubbles . ] + run & & ( @xmath72 k ) & ( @xmath73 k ) & ( @xmath74 ) & ( @xmath74 ) & ( @xmath74 ) & ( @xmath74 ) & ( @xmath74 ) & ( @xmath75 ) & ( @xmath75 ) + a & 8/7 & 5 & 3 & 2.2 & 0.2 & 3.5 & 2.1 & 7.7 & @xmath76 & @xmath77 + b & 8/7 & 5 & 3 & 2.2 & 0.1 & 3.6 & 2.1 & 7.6 & @xmath76 & @xmath77 + c & 10/4 & 3 & 3 & 0.7 & 0.4 & 1.9 & 0.9 & 4.7 & @xmath78 & @xmath79 + d & 8/8.5 & 5 & 3 & 2.7 & 0.2 & 4.4 & 2.8 & 9.2 & @xmath80 & @xmath81 + e & 8/8 & 10 & 3 & 2.5 & 0.2 & 3.9 & 2.5 & 8.4 & @xmath76 & @xmath82 + f & 8/7 & 5 & 3 & 2.7 & 0.2 & 5.2 & 2.4 & 9.4 & @xmath80 & @xmath83 + g & 8/6.5 & 7 & 5 & 4.5 & 0.7 & 5.7 & 4.6 & 15.2 & @xmath84 & @xmath85 + h & 8/7 & 4 & 3 & 2.0 & 0.3 & 3.3 & 2.0 & 7.2 & @xmath84 & @xmath86 + [ table2 ] we use zeus code ( @xcite ; @xcite ) and adopt 3-d cartesian coordinates . the advantage of choosing cartesian coordinates rather than spherical or cylindrical coordinates is that we can avoid the singularity on the polar axis arisen by one term of the viscous stress tensor . computational domain is from @xmath96.4 kpc to + 6.4 kpc in the x-,y - direction , and 0@xmath912 kpc in the z - direction . z - axis stretches along the galactic pole , and x - y plane is the galactic plane . sgr a * is located at the origin . we adopt non - uniform grid , with @xmath87 , @xmath88 , and @xmath89 . the numbers of meshes are i=128 , j=128 and k=120 in x-,y- , and z - direction respectively . we use the reflecting boundary condition on the lower boundary ( @xmath90 ) , and choose the outflow boundary condition on the other five boundary surface . we assume that the initial interstellar medium is an isothermal sphere in hydrostatic equilibrium state , i.e. , the gradient of the gas pressure balances the gravity . specifically , we assume the gravitational force given by stars and dark matter in a simplified form : ( r)=- , where @xmath91 . it will give a constant velocity dispersion of stars , and the velocity dispersion is 100 @xmath92 here . this is very similar to the circumstance in the galactic bulge . in the recent work by @xcite , a @xmath93-model was assumed to describe the gas density profile of galactic hole , and @xmath94 scales from @xmath95@xmath9@xmath96 @xmath97 at 1 kpc to @xmath98@xmath9@xmath99 @xmath97 at 10 kpc . the number density profile of electrons in our simulations is described by the form : n_e== , [ ne ] where @xmath100 is the average number of free electrons per nucleon , and @xmath101 for solar composition , @xmath102 is the atomic mass unit , @xmath46 is the electron number density at @xmath103 kpc , @xmath104/1kpc . the value of @xmath46 is @xmath95 @xmath97 in the `` basic run '' ( run a ) , and the density profile in 1@xmath910 kpc is well within the observational range mentioned above , while beyond 10 kpc , the gas has little effect on _ fermi _ bubbles . more realistic forms of gravity and gas distribution were adopted in @xcite and @xcite . the simplified form used here would not influence the results significantly , since the difference of density distribution between the two forms is not so large in the bulge or halo . the difference becomes significant close to the galactic plane but the _ fermi _ bubbles are far away from the galactic disk . the temperature of ism is @xmath105 k , which is determined by the velocity dispersion @xmath106 . the temperature is almost the same as that in @xcite . one important massive structure exists in the galactic center region , i.e. , the central molecular zone ( cmz , @xcite ) . it is elongated along the galactic plane , just surrounding sgr a * , with total mass of several @xmath107 . the length is 400 pc , and the height is 75 pc . as has been shown by @xcite , this structure has influence on the motion of winds from sgr a. it can collimate the winds to the perpendicular direction of the galactic plane . in the simulation , the cmz is set to be a torus - like structure located on the x - y plane , with inner radius of 80 pc and outer radius of 240 pc . it is in hydrodynamic equilibrium , and the rotating velocity is @xmath108 . the ratio between the height and radius is set to be 0.15 in all runs . from our test simulations , we find that the ratio does not influence the results significantly when it increases from 0.15 to 0.25 . the maximum thickness of cmz is 72 pc , close to observational result . the density of cmz is set to be a constant . the total mass of cmz is set to be @xmath109 . as mentioned in @xcite , cmz can not be blown away by the winds because the ram pressure force impacted on cmz is much smaller than the gravitational force . but the top and bottom parts of cmz can be affected by kelvin@xmath9helmholtz ( kh ) instability and will form an interesting structure , which can explain x - ray observations ( see 4.3 ) . the wind is injected from the inner boundary of the simulation , which has a height of 20 pc , and a width of 16 pc . the initial energy density of ism around inner boundary is @xmath110 . this pressure around the black hole supplies a threshold and only winds with ram pressure higher than this value will be able to push the ism away and induce shocks in the galactic halo . in most runs of our model , the ram pressure of wind is about twice the initial pressure of ism around the injection region . the hydrodynamic equations describing the interaction process are as follows . viscosity and thermal conductivity are included . @xmath111 @xmath112 @xmath113 @xmath114 here @xmath115 is the density of the gas , _ p _ is the gas pressure , _ e _ is the internal energy density of the gas , v * is the velocity , * t * is the viscous stress tensor , _ t _ is temperature , @xmath49 is the viscosity coefficient , @xmath116 is the heat conductivity coefficient , @xmath117 is the unite tensor . the relationship between the gas pressure and the internal energy density is described by @xmath118 . radiative cooling is neglected . we have estimated the total energy lost by bremsstrahlung cooling within 10 myr , and found that it is no more than a few percent of the total energy injected by wind . the values of viscosity coefficient @xmath49 adopted in our models are shown in table 1 . for comparison , we also run a model with @xmath119 ( run c ) . as argued in @xcite , the nature of viscosity is still highly uncertain . for a fully ionized , unmagnetized plasma , the dynamical viscosity coefficient is ( @xcite ) _ visc=6.010 ^ 3()^-1()^5/2 ^-1 ^ -1 [ equvisc ] where @xmath120 is the coulomb logarithm . the viscosity coefficient is 2 @xmath121 for a typical temperature of @xmath122 k in the shocked ism , and @xmath123 for @xmath72 k inside the bubble . in the present work , for simplicity we set the viscosity coefficient to be a constant which is very close to the value in the shocked ism while quite different from that inside the bubble . but in the cmz region we calculate the viscosity coefficient according to equation ( [ equvisc ] ) . cmz gas will not be suffered from the effect of viscosity since the viscosity coefficient is very low there . + as pointed out by @xcite , viscosity plays an important role because it can suppress instabilities so that we can get a smooth edge of the bubbles . the value of viscosity also influences the width of the bubble . we will discuss this point in more detail in 4.4 . we also include thermal conduction in our models except run e. it makes the distribution of gas inside bubbles uniform . the heat flux @xmath124 is given by : q =- t , here @xmath116 is the coefficient of thermal conductivity . for a fully ionized gas , @xmath116 is given by ( @xcite ) : 210 ^ -4 ^-1 k^-1 cm^-1.[thercon ] in reality , thermal conduction will be strongly affected by magnetic field . specifically , in the direction perpendicular to the magnetic field , thermal conduction will be strongly suppressed because it is difficult for the particles to move across the field lines . in addition , in a collisionless fluid , thermal conduction would be saturated , but the calculation of heat flux in this case is still on a phenomenological level with an artificially assumed factor ( @xcite ) . we find that our results are not sensitive to the value of @xmath116 . even a value of the coefficient of thermal conductivity orders of magnitude lower than that determined by eq . ( [ thercon ] ) is enough to smooth the distribution of gas within the bubble . in our model , we identify the cd to be the edge of the observed _ fermi _ bubbles . the region of shocked ism is the `` surrounding region '' . when magnetic field is included , the field lines in this region will be aligned with the cd , which prevents the diffusion of relativistic electrons across the bubble edge . this then explains why the edge of the bubbles is so sharp ( @xcite ) . this mechanism also applies to our model since in reality magnetic field should exist . if we only want to explain the morphology of the bubbles , we find that we have relatively large freedom in terms of the values of velocity and mass flux of winds . for example , we can use a smaller wind velocity and a higher mass outflow rate ( see model g , or fig . [ label8 ] ) , or a larger wind velocity and a lower mass outflow rate to get the `` correct '' morphology . however , observations put additional constraints , such as temperature . we choose run a as our `` basic run '' because not only the morphology but also other properties of the bubbles are consistent with observations . in run a , winds need to last for 12.3 myr to get the `` correct '' morphology of the bubbles , as shown by fig . [ label1 ] . the height and width of the bubble are 8 kpc and 7 kpc respectively , which corresponding to a projected bubble with a latitude of @xmath13 and a longitude of @xmath13 . although the wind is set to be injected isotropically , the massive cmz surrounding sgr a * blocks the lateral movement of winds , and forces them to move upwards . in other words , the cmz successfully collimates the wind to the perpendicular direction of the galactic plane . this is why we can obtain a bubble with narrow waist near the galactic plane , instead of a hemispherical bubble buckling on the galactic plane . we have tried to explore when the activity of sgr a * quenched and entered into the quiescent state by running `` run b '' . we find that only if the quiescent time is shorter than 0.2 myr , the result will not be affected , i.e. , showing a significant conical structure in x - ray band with latitude @xmath125 ( see lower panels in fig . [ label3 ] ) . because the quiescent timescale is so short compared with the age of _ fermi _ bubbles , although in all other runs in this work sgr a * do not enter into the quiescent state as it should be , our simulation results will not be affected . for run a , the total energy injected by sgr a * is about @xmath126 erg . this energy is comparable to the injected energy from galactic center estimated from some observations to some structures , such as the nps structure , gcl , and emr ( @xcite ) . the total internal energy of _ fermi _ bubbles in our simulation is @xmath127 erg , which is consistent with the observational value . the total kinetic energy of the _ fermi _ bubbles is only @xmath128 erg , much smaller than the internal energy . this is because the speed of the gas inside the bubble is subsonic . the total mass inside the bubble is about a few times @xmath129 , which is much lower than the estimation of @xmath130 based on the assumed upper - limit of an average density @xmath131 @xmath97 in @xcite . this is because the density in our simulation is about three orders of magnitudes lower than the assumed value in @xcite , as shown by fig . [ plot2 ] . we have the following comments to this `` discrepancy '' . firstly , our observational constraint on the density is poor , thus the total mass of @xmath130 is subject to a large uncertainty . secondly , since the coefficient of thermal conductivity adopted in our work is low , we may have underestimated the evaporation process which may play an important role in transporting mass from the surrounding gas , including cmz and shocked ism , into the bubbles . thirdly , we assume the ism homogeneous for simplicity , while in reality the ism is likely to be clumpy . the dense clouds may be difficult to be blown away by winds so they will stay inside the bubble , which will significantly increase the mass of the gas within the bubbles . the temperature of the gas is determined by the following equation : t= , where @xmath132 is temperature , @xmath49 is the molecular weight , which is 0.61 for solar composition . temperatures of different runs are given in table 2 . in general , the temperature inside the bubble is several times @xmath72 k. although there are some new results of milky way s hot halo recently , the temperature inside the _ fermi _ bubbles is still lack of data . so our result can be regarded as a prediction . we will discuss the temperature in the surrounding region between the cd and the forward shock in 4.3 . we have calculated the predicted x - ray image by considering the bremsstrahlung radiation . [ label3 ] shows the result . we can see that the morphology is consistent with the limb - brightened x - ray structure obtained in _ rosat _ observations ( @xcite ; @xcite ) . this structure corresponds to the shocked ism . here , we only calculate the bremsstrahlung radiation , so our images are only for qualitative comparison . the temperature inside the bubble is two orders of magnitude higher than the surrounding region while the density is two orders of magnitude lower . therefore the outer region is much brighter than the interior of the bubbles in x - ray band . recent _ suzaku _ observations have revealed that the temperature of the surrounding region in high latitude ( @xmath133 ) is around 0.3 kev and the thermal pressure is @xmath134 dyn @xmath135 ( @xcite ) . our model is in good consistency with their results . in run a , the temperature and thermal pressure in the same location are 0.4 kev and @xmath136 dyn @xmath135 respectively ( also see table 2 , note that the temperature is space - averaged value of the shocked ism ) . in contrast , in jet model and quasar outflow model , the predicted temperature is larger than a few kev ( @xcite ; @xcite ; @xcite ) and 1 kev ( @xcite ) respectively . we think that the main reason for such a discrepancy is that the wind velocities , and more importantly the mass fluxes of wind , in these two models are too high . we have calculated the bremsstrahlung radiation , and found that the total lost energy is no more than ten percent of the total internal energy of the surrounding region . so cooling effect is very weak . we note that the interaction region between the winds and the cmz gas is also quite bright in x - ray band , which looks like a cone upside down on the galactic plane . this structure explains the features observed in the _ rosat _ x - ray survey by @xcite and @xcite . as we have mentioned in 4.1 , the brightness of the conical x - ray structure near the gc observed in [email protected] kev is related with the time duration starting from the quenching of the past activity of sgr a. from run b , we find that the time duration should be no more than @xmath11 0.2 myr . observations also show that we can only see the east ( left ) x - ray structure of the northern sky . together with the bending of the northern _ fermi _ bubble , we speculate that this phenomenon may be caused by the galactic wind blowing from the east to the west in the northern sky . hence , the forward shock in the east will be stronger than the west , and both the temperature and the density of the shocked ism in the east will be larger than the west , inducing the asymmetric structure of x - ray emission . another possibility is that the initial ism is not symmetric , with the density in the left ( east ) part being higher . the _ rosat _ x - ray structure looks like an x - ray cavity . such kinds of cavities have also been observed in other galaxies or galaxy clusters . usually people think they are formed by the interaction between jets and ism or intergalactic medium . however our result reminds us that these cavities may be well formed by the interaction between the winds ( rather than jet ! ) from the central agn and the igm . as pointed out by @xcite and @xcite in the case of the cavity in m87 , if the cavity were formed by the jet , we would expect a sharp bow shock regions between a jet and surrounding medium . this structure has never been observed . the morphology also depends on the viscosity coefficient @xmath49 , as shown by fig . [ label4 ] . we can see from the figure that if the viscosity coefficient is larger , the bubble will be more spherical . the winds near the cmz suffer from the viscous force because of the large velocity gradient on the @xmath137@xmath9@xmath138 plane , and they are slowed down by the cmz significantly . if viscosity coefficient is larger , the kinetic energy of wind gas will be dissipated into internal energy more efficiently , then the thermal pressure close to the gc will be larger . therefore , the opening angle of the blown - up cmz gas will be wider , which causes the bubble more spherical . viscosity can suppress both the rayleigh - taylor ( rt ) and kelvin - helmholtz ( kh ) instabilities . following equations ( 18 ) and ( 19 ) from @xcite , we can estimate the timescales for the growth of both instabilities . for example , for the rt instability to form a @xmath139 kpc structure at height @xmath140 kpc , the required timescale is about 5 myr ; while for the kh instability , it is about 1.5 myr . so both instabilities can grow up during the formation of the _ fermi _ bubbles . from the left plot of fig . [ label4 ] we can see that , when viscosity is not included , large rolls with typical length scales of @xmath11 kpc inside the shocked ism are formed . however , for all the other runs with viscosity included , no such rolls are found . another role of viscosity is the viscous heating . this effect is important in the interaction region between winds and cmz . in this region , the main components of viscous stress tensor are @xmath141 @xmath142 and @xmath143 @xmath144 , ~t_xz + t_yz . for the interaction region , @xmath145 . replacing @xmath146 with @xmath147 and @xmath148 with @xmath149 pc , we can estimate that the time - scale for the winds to pass through this region is @xmath11 1 myr . the density of the cmz gas blown up in the interaction region is @xmath150 . then the increase of temperature is @xmath151 k. what is the role of thermal conduction ? we study this problem in run e ( without thermal conduction ) . [ label6 ] shows the distribution of the electron number density and temperature . a jet - like structure along the @xmath152-axis is clearly seen . since the grids along the @xmath152-axis are elongated , we need to check whether this feature is artificial or not . we have done such a test and found that this feature is likely real . the temperature of this structure is relatively low but the density is high . their formation mechanism is as follows . in the inner region , the massive cmz gas acts like a wall around sgr a , preventing the winds from expanding in the horizontal direction . the winds collide with the cmz and the kinetic energy of winds is converted into thermal energy , thus the temperature and pressure increase . the high - pressure gas then escapes towards the polar direction , squeezing the wind from sgr a * and causing the formation of this jet - like structure . however , when we include thermal conduction as in most of our runs , this structure disappears . this is because thermal conduction can efficiently transport energy between the regions with different temperature thus smooth out this structure . for simplification , we have adopted a power - law distribution for the density distribution of the initial ism : @xmath153 , where _ a _ and _ n _ are constants , and @xmath154 is in unit of @xmath155 pc in our simulations . the value of _ a _ has a weak influence on the age of the _ fermi _ bubbles . this can be seen in run f ( refer to fig . [ label7 ] ) . although the density is two times higher than the basic run , it only takes 20% more time to form the bubbles . this is easy to understand from equations ( [ equ_r2 ] ) and ( [ equ_rfan ] ) . different from the parameter _ a _ , the value of _ n _ is more important to influence the evolution of the bubbles . for example , our simulations show that if the index _ n _ is changed into 2.0 from 1.6 while _ a _ remains unchanged compared to run f , the age of the bubble would be 7 myr , which is half of the age of run f. the temperature of the shocked ism in latitude @xmath133 is about 1 kev in this case . physically , this is because the kinetic energy of shocked ism is nearly a constant fraction of the total energy injected from gc . specially , from equations ( [ equ_r2 ] ) and ( [ equ_rfan ] ) , we can obtain : m_c^2_2=2a p_wt,[energyequ ] where @xmath156 and @xmath157 are the mass and velocity of shocked ism respectively . we find that @xmath158 is almost a constant : @xmath159 . so we can approximately rewrite the righthand side of equation ( [ energyequ ] ) as : @xmath160 ( @xmath161 is a constant ) , or @xmath162 when @xmath40 keeps unchanged . for the two cases mentioned above , if _ a _ is doubled , @xmath156 is doubled , while _ n _ changed from 1.6 to 2.0 , @xmath156 is only 1/8 of the former . that s why the ages are 20% larger and one - half smaller respectively . in run g , the mass flux of winds is three times higher than run a , while the wind velocity is the same . the age is 34% shorter than run a. this is very close to the result of 31% by the simple analytical analysis shown by equation ( [ equ_r2 ] ) . because of the increase of the wind power , the velocity of the forward shock increases thus the temperature of the shocked ism becomes higher , while the increase of temperature inside the bubble is not so obvious . in run h , we reduce the wind velocity but keep the kinetic power of wind ( @xmath40 ) unchanged compared to run a. we find that the kinetic and thermal energy and temperature of different regions do not change much ( refer to table [ table2 ] ) . this means that , the results mainly depend on the kinetic power , while the velocity and mass outflow rate are degenerate . this is also easy to be understood from equations ( [ equ_r2 ] ) and ( [ equ_p ] ) . the gas inside the bubbles mainly comes from the blown - up cmz gas instead of winds injected from sgr a. in addition , the evaporation of the bubble edge , which is determined by thermal conduction , should also supply additional gas . but since there is large uncertainty on thermal conduction coefficient , it is hard for our model to predict the exact mass of the gas inside the bubbles . in addition to the total mass , another interesting quantity is the temperature of the gas inside the bubbles . to estimate the temperature , we need to know the pressure . this quantity is equal to the gas pressure of the shocked ism , which is well determined by the density of ism and the wind power from equation ( 10 ) . unfortunately , the uncertainty of the gas density inside the bubbles mentioned - above makes it hard to precisely predict the temperature . based on our simulation of run a , we can only estimate the temperature of gas within the bubbles to be in the range of @xmath163 k. observational constraints are still lacking . we have performed hydrodynamical numerical simulations to study the formation mechanism of the _ fermi _ bubbles detected by _ fermi_-lat . our main aim in the present paper is to explain the morphology and the thermodynamical properties of the bubble , but leaving the study of the production of @xmath17-ray photons and the explanation of the spectrum to our next work . while sgr a is quite dim at the present stage , many observational evidences indicate that this source should be much more active in the past . specifically , one possibility suggested by a previous work is that the mass accretion rate of the hot accretion flow in sgr a * should be @xmath0 times higher than the present value and this activity lasts for several myr ( totani 2006 ) . based on this scenario , we show that the observed _ fermi _ bubbles can be well formed by the interaction between the winds launched from the `` past '' hot accretion flow and ism . in our model , the winds last for @xmath24 yr and the activity of sgr a * was quenched no more than @xmath164 myr ago . the properties of wind such as the mass flux and velocity are not free parameters but obtained from the previous works on mhd numerical simulations of hot accretion flows . viscosity and thermal conduction are included which can suppress various instabilities and make the gas inside the bubble uniform . the required power of the winds is @xmath165 , which is fully consistent with the previous studies on the past activity of sgr a. the edge of the bubbles corresponds to the contact discontinuity which is the boundary between the shocked interstellar medium and the shocked winds . properties of the bubbles such as the morphology and the total energy are consistent with observations . the limb - brightened _ rosat _ x - ray structure can be interpreted by the shocked ism behind the forward shock , while the conical - like x - ray structure close to galactic center is interpreted by the interaction region of wind gas and cmz gas . our model can also quantitatively explain both the thermal pressure and the temperature of the x - ray structure in high latitude position ( @xmath133 ) revealed by the recent _ suzaku _ observations . in addition to winds , jets should also co - exist with hot accretion flow ( yuan & narayan 2014 ) . in our model , we do not include the jet . we assume that the interaction between jet and the interstellar medium is negligible because , by definition , jet must be well - collimated and be as fast as the light . in this case , we expect that the jet will simply drill through the ism , with almost no interaction with the ism in the galaxy . we have also calculated the energy transformation efficiency in our model . we find that at @xmath166 kpc , @xmath11 60% of the total energy of winds injected from sgr a is transported into the ism . obviously , such a high efficiency is because of the large opening angle of winds . this result suggests that we may consider the role of winds in solving the cooling flow problem in some elliptical galaxies and galaxy clusters . usually people consider the heating of ism or intracluster medium by jets ( see , e.g. , @xcite and references therein ) . however , numerical simulations have found that jet may only be able to deposit their energy at @xmath167 kpc thus not very efficient ( @xcite ) . some solutions have been suggested , e.g. , the precession of a jet , or motions of intracluster medium ( see @xcite and @xcite ) . but another possible way is to invoke winds whose existence has been firmly established by both observational and theoretical studies . given our successful explanation of the formation of the _ fermi _ bubbles by the wind model , it is also worthwhile to study whether the x - ray cavities observed in galaxy clusters ( e.g. , @xcite ) , which have the similar morphology with the _ fermi _ bubbles , can be produced by winds . we thank fuguo xie , zhaoming gan , tomohisa kawashima , chao liu , hui li , and daniel wang for helpful discussions . we also thank the referee for constructive suggestions . this work was supported in part by the natural science foundation of china ( grants 11103061 , 11133005 , 11121062 , and 11103059 ) , the national basic research program of china ( 973 program , grant 2014cb845800 ) , and the strategic priority research program `` the emergence of cosmological structures '' of the chinese academy of sciences ( grant xdb09000000 ) . m. y. sun acknowledges support from the china scholarship council ( no . [ 2013]3009 ) . support for the work of m.s . was provided by nasa through einstein postdoctoral fellowship grant number pf2 - 130102 awarded by the chandra x - ray center , which is operated by the smithsonian astrophysical observatory for nasa under contract nas8 - 03060 . the simulations were carried out at shao super computing platform . , g. , morris . m. r. , terrier , r. , & goldwurm , a. 2013 , in astrophysics and space science proceedings , vol . 34 , cosmic rays in star - forming environments , ed . d. f. torres & o. reimer ( heidelberg : springer ) , 331	a pair of giant gamma - ray bubbles have been revealed by the _ fermi _ lat . in this paper we investigate their formation mechanism . observations have indicated that the activity of the supermassive black hole located at the galactic center , sgr a * , was much stronger than the present time . specifically , one possibility is that while sgr a * was also in the hot accretion regime , the accretion rate should be @xmath0 times higher during the past @xmath1 yr . on the other hand , recent mhd numerical simulations of hot accretion flows have unambiguously shown the existence of strong winds and obtained their properties . based on these knowledge , by performing three - dimensional hydrodynamical simulations , we show in this paper that the fermi bubbles could be inflated by winds launched from the `` past '' hot accretion flow in sgr a. in our model , the active phase of sgr a is required to last for about 10 million years and it was quenched no more than 0.2 million years ago . the central molecular zone ( cmz ) is included and it collimates the wind orientation towards the galactic poles . viscosity suppresses the rayleigh - taylor and kelvin - helmholtz instabilities and results in the smoothness of the bubble edge . the main observational features of the bubbles can be well explained . specifically , the _ rosat _ x - ray features are interpreted by the shocked interstellar medium and the interaction region between winds and cmz gas . the thermal pressure and temperature obtained in our model are in good consistency with the recent _ suzaku _ observations .
in recent years , quantum dot ( qd ) systems , as a potential artificial thermoelectric material , have attracted much attention and their thermoelectric properties have been widely studied both experimentally @xcite and theoretically@xcite . due to deviation from the wiedemann - franz law and suppression of the phonon contributions , thermoelectric efficiency in qd systems can be significantly enhanced@xcite . on the other hand , many interesting thermoelectric effects have been observed or predicted in single qd and double quantum dots ( dqds ) systems in the coulomb blockade ( cb ) regime@xcite and kondo regime@xcite . for example , the thermopower and the thermal conductance as a function of gate voltage present cb oscillations@xcite . the transient thermopower of a single qd system can be strongly enhanced and modified under a time - dependent gate voltage@xcite . in a single qd with multilevels@xcite or a serial - coupled dqds@xcite , the thermal conductance can be strongly enhanced at high temperature due to the bipolar effects , leading to strong suppression of thermoelectric efficiency . when interference effects are considered in parallel coupled dqds system , the thermoelectric efficiency can be significantly enhanced by fano effect and coulomb interaction@xcite . in particular , the strong dependence of thermopower on temperature and gate voltage provide a sensitive way to detect and understand the kondo correlations@xcite . the recent experimental observation of the spin seebeck effect@xcite inspires the research on spin effects in energy transport in qd systems@xcite and opens a new view of thermal manipulation of spintronic devices and the design of spin current generator . spin - dependent thermoelectric effects in the single qd and dqds system connected to two ferromagnetic leads with collinear or noncollinear magnetic configurations is widely studied in the linear and nonlinear response regime@xcite . the results indicate that the thermoelectric coefficients are strongly dependent on the spin polarization , the relative magnetic configurations of ferromagnetic leads and the asymmetry of dot - lead coupling . as mentioned by jzef barna@xcite , due to long spin - relaxation time , the spin accumulation in leads induces spin - dependent splitting of the chemical potential in the leads . as a result , the spin efficiency can be larger than the charge efficiency and a pure spin thermopower can be obtained@xcite . up to now , most earlier theoretical works about spin effects in heat transport mainly focused on single qd or parallel coupled dqds system . more recently , people start to study thermoelectric effects in a serial - coupled dqds system coupled to ferromagnetic leads@xcite . even so , still much more work is needed to understand interplay between spin effect and heat transport in serial - coupled dqds system , particularly in the presence of the magnetic field or the spin accumulation in leads . in the present work , we consider a lateral dqds with spin - dependent interdot coupling connected to two ferromagnetic electrodes . spin - dependent interdot coupling can be achieved by applying a static magnetic field on the tunneling junction between two qds and induces the levels spin - dependent splitting . the tunable system properties further enrich the thermoelectronic and thermo - spin transport properties in dqds system . by using the nonequilibrium green s function technique , thermoelectric and thermo - spin coefficients are obtained . it is found that in parallel configuration the thermoelectric efficiency can achieve considerable values around spin - down resonance levels when the effective interdot coupling and tunnel coupling between qds and the leads for spin - down electrons are small . on the other hand , in the presence of the magnetic field , the spin accumulation in leads strongly suppresses the thermoelectric efficiency . in particular , the thermoelectric and thermo - spin efficiency are strongly enhanced by the intradot coulomb interactions and can reach very high value at an appropriate temperature . moreover , a pure spin thermopower can also be obtained . the rest of this work is organized as follows . in sec.ii , we introduce the model of lateral dqds system with spin - dependent interdot coupling and derive the basic analytical formulas . in sec.iii , we present the corresponding numerical results about the influence of the related parameters on the thermoelectric and thermospin properties in the present device . finally , in sec.iv we summarize the work . we consider a system of lateral dqds with spin - dependent interdot coupling and two ferromagnetic leads . the magnetizations of the leads are assumed to be collinear , and can be either in the parallel or antiparallel configuration . the hamiltonian describing the system can be decomposed into three parts , @xmath0 , @xmath1 . the first term describes the isolated dqds and is given by,@xmath2 \label{1}%\end{aligned}\ ] ] where the operator @xmath3 @xmath4 creates ( annihilates ) an electron with energy @xmath5 and spin @xmath6 in qd @xmath7 . we assume that the level spacing is very large so that we can only consider a single energy level in each qd . @xmath8 is intradot coulomb interaction constant and @xmath9 is the particle number operator . the last term sum in eq.(1 ) describes the spin - dependent tunneling between two qds , where @xmath10 is defined as : @xmath11 for @xmath12 or @xmath13 for @xmath14 . this spin - dependent tunneling strength can be induced by a static magnetic field applied to the barrier region between the two qds . when electron passes through the potential barrier region , the electron with spin parallel to the magnetic field have a larger tunneling probability than one with spin antiparallel to the magnetic field due to larmor precession of the electron spin inside the potential barrier@xcite . the second term @xmath15 of hamiltonian describes the noninteracting electrons in the leads and is written as @xmath16 , where @xmath17 creates ( annihilates ) an conduction electron with energy @xmath18 , momentum @xmath19 and spin @xmath6 in the @xmath20 electrode . the third term of hamiltonian describes the coupling between the dqds and the electrodes and reads @xmath21 , where @xmath22 is the hopping matrix - element between the qd @xmath7 and the electrode @xmath20 , which is assumed to be independent of momentum @xmath19 . in the following calculation , we define the line - width matrix as @xmath23and @xmath24 and in a wide - band limit , @xmath25 is independent of the energy . using the matrix representation , we have@xmath26{cc}% \gamma_{11l}^{\sigma } & \gamma_{12l}^{\sigma}\\ \gamma_{21l}^{\sigma } & \gamma_{22l}^{\sigma}% \end{array } \right ) = \left ( \begin{array } [ c]{cc}% \gamma_{1l}^{\sigma } & 0\\ 0 & 0 \end{array } \right ) \nonumber\\ \gamma_{r}^{\sigma } & = \left ( \begin{array } [ c]{cc}% \gamma_{11r}^{\sigma } & \gamma_{12r}^{\sigma}\\ \gamma_{21r}^{\sigma } & \gamma_{22r}^{\sigma}% \end{array } \right ) = \left ( \begin{array } [ c]{cc}% 0 & 0\\ 0 & \gamma_{2r}^{\sigma}% \end{array } \right ) \label{2}%\end{aligned}\ ] ] with @xmath27 and @xmath28 . when the spin polarization of the leads is considered , the matrix - element may be assumed in the form : @xmath29 and @xmath30 with the sign @xmath31 ( @xmath32 ) for parallel ( antiparallel ) alignment of the magnetization in two electrodes . @xmath33 is related to the lead s spin polarization . @xmath34 describes the tunneling coupling between the dqd and the @xmath20-th electrode for @xmath35 . in the following calculation we assume @xmath36 , and @xmath37 . by using nonequilibrium green s function technique , the electronic and heat current from the @xmath20 ( @xmath38 , @xmath39 ) lead flowing into the qds can be written in the forms@xcite:@xmath40 t_{\sigma}\left ( \omega\right ) \label{3}%\]]@xmath41 t_{\sigma}\left ( \omega\right ) \label{4}%\ ] ] where @xmath42 , @xmath43^{-1}$ ] is the fermi distribution function for the @xmath20 electrode with chemical potential @xmath44 , temperature @xmath45 and boltzmann constant @xmath46 . @xmath47 is the transmission coefficient corresponding to spin channel @xmath6 and is given by@xmath48\ ] ] here @xmath49 and @xmath50 are , respectively , the retarded and advanced green s function in the dqd in @xmath51 space , with @xmath52 ^{\dagger}$ ] . employing the equation of motion and adopting the hartree - fock truncating approximation for higher order green s functions , the retarded green s function can be obtained as from the dyson equations@xmath53{cc}% g_{11\sigma}^{r}(\omega ) & g_{12\sigma}^{r}(\omega)\\ g_{21\sigma}^{r}(\omega ) & g_{22\sigma}^{r}(\omega ) \end{array } \right ) = g_{\sigma}^{r}(\omega)+g_{\sigma}^{r}(\omega)\sigma_{\sigma}% ^{r}g_{\sigma}^{r}(\omega)\ ] ] where @xmath54 $ ] is the retarded self - energy in the wide - band limit and @xmath55 is the retarded green s function for the isolated dqd ( without coupling to electrodes ) which is expressed as @xmath56{cc}% c_{1\sigma } & t+\delta_{\sigma}\delta t\\ t+\delta_{\sigma}\delta t & c_{2\sigma}% \end{array } \right ] ^{-1 } \label{7}%\ ] ] where @xmath57 and @xmath58 , the averaged occupation number of electrons with spin @xmath6 in the @xmath7-th quantum dot , can be calculated self - consistently by using the formula @xmath59 , where @xmath60 is the lesser green function in @xmath51 space and can be obtained by the keldysh equations @xmath61 with the lesser self - energy @xmath62 $ ] . it should be pointed that the hartree - fock decoupling approximation scheme is reasonable when the temperature during our calculation is higher than the kondo temperature . in the linear response regime , the spin accumulation in electrodes can be neglected when spin relaxation time is sufficiently short@xcite . in this case , the voltage induced by temperature gradient can be treated as independent of spin . the electric and heat currents through the system can be written in the form@xmath63@xmath64 with @xmath65 . in the following calculation we assume the chemical potential @xmath66 and temperature @xmath67 , thus the fermi distribution function @xmath68 . the electronic and thermal conductances can be calculated by the formula @xmath69 and @xmath70,$ ] respectively . the thermopower is defined as thermoelectric voltage @xmath71 induced by a temperature gradient @xmath72 under the condition of @xmath73 , _ _ i.e.__@xmath74 @xmath75 . the efficiency of heat - electricity conversion of the system is described by the dimensionless figure of merit ( fom ) @xmath76 , where @xmath77 denotes the thermal conductance corresponding to the phonon contribution , which is suppressed in such dqd - electrodes structure , and is ignored in the present paper . if spin relaxation time of systems is sufficiently long , the spin accumulation in electrodes should be considered@xcite , which leads to that the chemical potential difference @xmath78 between the two leads is related to the spin channel . thus , in this system one can define a spin bias @xmath79 induced by temperature gradient via the formula @xmath80 . corresponding spin - resolved electric and heat currents can be expressed as @xmath81@xmath82 \label{11}%\ ] ] with @xmath83 @xmath84 . the charge and spin conductance are calculated as @xmath85 $ ] and @xmath86 $ ] , respectively , while thermal conductance is given by @xmath87 $ ] , which is identical to the value without spin accumulation in such a system . one may introduce from eq.(10 ) the spin - dependent thermopower @xmath88 in spin channel @xmath6 under the condition of current @xmath89 vanishing@xcite . then the charge and spin thermopower are defined as @xmath90 $ ] and @xmath91 , $ ] respectively . corresponding charge and spin figure of merit are defined as @xmath92 and @xmath93 , respectively , which describe heat - to - charge - voltage and heat - to - spin - voltage conversion efficiency of the present system when spin accumulation is considered . @xmath94 has the form of absolute value because the spin conductance may be negative in some regions . we now numerically study spin - dependent thermoelectric effects in a dqds system based on the above formula , which will be presented according to three premises : without spin accumulation in the leads and @xmath95 ; without spin accumulation in the leads and @xmath96 ; with spin accumulation in the leads . for the sake of simplicity , we assume two dots have same bare energy levels @xmath5 and are spin degenerate , i.e. @xmath5 @xmath97 , which can be tuned by gate voltage in experiment . the relevant parameters are chosen as : @xmath98 , @xmath99 , @xmath100 mev , @xmath101 mev . consider the case that the magnetizations of the leads are in parallel configuration , there is no spin accumulation in the leads and the intradot coulomb interactions are ignored . the tunnel coupling between two qds is spin - dependent due to the external magnetic field , such a coupling dqds can be equivalent to a single qd with four spin - dependent levels , @xmath102 , @xmath103 , @xmath104 , and @xmath105 , which corresponds to poles of the green function @xmath106 given by eq.(7 ) . these levels are identified as @xmath107 ( u=1,2 ) occupied by spin - up electrons and @xmath108 ( d=3,4 ) occupied by spin - down electrons , and satisfy @xmath109 for chosen parameters . the influence of splitting @xmath110 of interdot coupling on the thermoelectric transport is shown in fig . 1 . four resonance peaks appear in the electrical conductance spectrum when the spin - dependent level crosses the fermi level ( @xmath111 ) , see fig . it can be clearly seen that the variety of @xmath110 only changes the electrical conductance s resonance positions and do not influence its magnitude . it is found that this hold true for the thermal conductance @xmath112 at low temperature . however , the behavior of @xmath112 is more complicated at relatively high temperature ( shown in fig . 1(b ) ) , and the thermopower @xmath113 and the figure of merit @xmath114 are more sensitive to the variety of @xmath110 . the antisymmetry , oscillation and sign reversal of @xmath113 also are obtained from fig.1(c ) . as shown in fig . 1(d ) , @xmath114 presents typical double - peak structure . when @xmath110 increases , for the two peaks of @xmath114 located around @xmath115 , not only their magnitudes augment but also their degree of asymmetry increases . in particular , @xmath114 exhibits a behavior similar to crossover and can reach very high value when @xmath116is close to @xmath117 ( see the inset of fig . 1(d ) ) . this is because when @xmath116is close to @xmath117 , the levels @xmath118 and @xmath119 get much closer to each other and the spin - down resonance channels are significantly blockaded due to the strong reduce of the interdot tunneling for spin - down electrons . while with @xmath110 increasing the side peaks located at @xmath120 only shift their positions with unchanged the maximum . next , we study the influence of spin polarization @xmath121 on the thermoelectric transport . the relevant thermoelectric quantities versus dot s level position @xmath122 for different spin polarization @xmath121 are presented in fig . when @xmath121 increases , the electrical and thermal conductance peaks ( the thermopower peaks ) located at @xmath123 increase ( decreases ) while the electrical and thermal ones ( the thermopower ones ) located at @xmath115 decrease ( increases ) because @xmath124 , @xmath125 increases and @xmath126 , @xmath127 decreases . these properties of @xmath128 , @xmath112 and @xmath113 determine the behavior of @xmath114 . as shown in fig . 2(d ) , with increasing @xmath121 the @xmath114 near @xmath123 is suppressed while the @xmath114 near @xmath115 is intensively enhanced . the dependence of @xmath114 on spin polarization in the leads and on spin - dependent levels may be useful for designing spin - dependent thermoelectric devices . however , it should be noted that when the spin polarization in leads is very high , for example , @xmath129 , the tunneling rate for spin - down electrons is close to zero leading to significant suppression of transport through the spin - down levels and thus the corresponding @xmath114 peaks rapidly decrease with increase of @xmath121 , which can be clearly seen from the inset ( 2 ) of fig . 2(d ) . when the magnetic configuration changes from parallel to antiparallel , the electrical conductances are suppressed for four resonance channels , but the changes of the thermal conductance @xmath112 and the thermopower @xmath113 are dependent of the resonance levels ( see fig . 3 ) . @xmath112 ( @xmath113 ) lowers ( heightens ) for the spin - up resonance channels and heightens ( lowers ) for the spin - down resonance channels . as a result , @xmath114 in antiparallel configuration presents higher peaks around the spin - up resonance channels and lower peaks around spin - down resonance channels comparing with in parallel configuration . for antiparallel configuration , the tunneling rates between the leads and the qds satisfy @xmath130 , which leads to the transmission coefficient @xmath131 is spin - independent . then the thermoelectric quantities have the same intensities at different resonance points . although the electric and thermal conductance experience an decreasing trend with increasing @xmath121 , the thermopower and the figure of merit are independent of @xmath121 . based on the discussions above , in order to obtain high thermoelectric efficiency one should tune the resonances points to spin - down levels and take @xmath132 in parallel configuration for fixed temperature and splitting of interdot coupling . now we focus on the case taking into account the intradot coulomb interactions . under the hartree - fock approximation , such a coupling dqds system can be equivalent to a single qd with eight energy levels , in which four levels are occupied by spin - up electrons and other four are occupied by spin - down electrons@xcite . to determine the eight spin - dependent effective energy levels one need self - consistently solve the equation @xmath133^{-1}\right\vert = 0 $ ] due to @xmath96 . 4 presents the relevant thermoelectric quantities versus the dots level position @xmath122 with different temperature and @xmath134 . at low temperature eight resonance peaks of @xmath128 and @xmath112 can be clearly seen from fig . 4(a ) and ( b ) . the peaks @xmath135 , @xmath136 , @xmath137 , @xmath138 ( @xmath139 , @xmath140 , @xmath141 , @xmath142 ) correspond to the electron transport through spin - up ( down ) resonance channels . with the temperature increasing the conductance peaks become wider and lower due to broadening of the fermi distribution function . more interestingly , at high temperature , the thermal conductance presents five high peaks due to the bipolar effects@xcite . @xmath113 and @xmath114 experience a nonmonotonous variation trend with increasing temperature ( see fig . 1(c ) , ( d ) ) . in particular , at high temperature , two high peaks of @xmath114 appear near symmetry point @xmath143 , resulting from a suppression of @xmath112 and an enhancement of @xmath113 . comparing with the case in the absence @xmath8 , it can be found that coulomb interactions shift the positions of the peaks of @xmath114 and strongly enhance @xmath114 . as shown in the insert of fig . 4(d ) , the maximum of @xmath114 increases with coulomb interactions increasing and can reach very high values above 30 at an appropriate temperature , which is very useful for the thermoelectric energy conversion devices . it should be mentioned that with varying the relative magnetic configurations and the spin polarization of leads , the relevant thermoelectric quantities exhibit same variation trend as that without coulomb interactions except for peaks around @xmath143 , which are unaffected . for the systems with long spin relaxation time , the spin accumulation in electrodes should be considered@xcite . 5 presents the thermoelectric coefficients versus @xmath122 in the presence of spin accumulation . it is found by comparing fig.(3 ) with fig.(5 ) that the spin accumulation in electrodes has no influences on the electrical conductance and the thermal conductivity ( not shown ) but strongly influences on the thermopower and the figure of merit . the peaks of @xmath144 and @xmath145 are much lower than the appropriate peaks in @xmath113 and @xmath114 . these properties are quite different from those in the absence of external magnetic field , i.e. the splitting of interdot coupling @xmath146 . since the levels are strongly spin - dependent in such system , the spin fom experiences the same variation trend as charge fom with varying the relative magnetic configurations and the spin polarization of leads , and the behaviors are similar to the case without spin accumulation discussed above . some novel properties also appear , for instance , in the absence of coulomb interactions , @xmath147 at resonance points @xmath115 while @xmath148 at resonance points @xmath120 ( see fig . 5(a ) ) . more interestingly , for certain @xmath122 ( pointed by arrow and being around @xmath149 in fig . 5(b ) ) , the charge thermopower vanishes while the spin thermopower can be finite because the thermopowers for different spin have same value but opposite sign , which means that a pure spin current without charge current may be produced by a temperature gradient in our system . this feature can also be obtained in the spectrum of figure of merit in fig . when the coulomb interaction is considered , near the symmetry point @xmath145 exhibits two very high peaks while @xmath94 completely vanishes . this is because the thermopower for different spin have the same value and sign leading to the enhancement of charge thermopower and the suppression of the spin thermopower . even so , @xmath145 and @xmath94 are strongly enhanced by the coulomb interactions , and @xmath94 can be larger than @xmath145 for selected dots level positions in coulomb blockade regime . this implies our proposed system is a good heat - to - spin - voltage converter . in summary , we studied spin effects in thermoelectric transport through a lateral dqds system with spin dependent interdot coupling and ferromagnetic electrodes . it is found that in parallel configuration the thermoelectric efficiency can achieve considerable values around spin - down resonance levels when the effective interdot coupling and tunnel coupling between qds and the leads for spin - down electrons are small . on the other hand , in the presence of the magnetic field , the spin accumulation in leads strongly suppresses the thermoelectric efficiency . the thermoelectric and thermo - spin efficiency are strongly enhanced by the intradot coulomb interactions and can reach very high value at an appropriate temperature . moreover , in such a dqds system a pure spin thermopower can be obtained . this work was supported by national nature science foundation of china ( grant nos . 10974124 and grant nos . 11004124 ) and shanxi nature science foundation of china ( grant no . 2009011001 - 1 ) . figure caption : fig . 1 ( color online ) thermoelectric coefficients ( a ) @xmath128 , ( b ) @xmath112 , ( c ) @xmath113 , and ( d ) @xmath114 as a function of dot s level position @xmath122 for varied splitting @xmath110 in parallel configuration with @xmath150 mev . @xmath110 is measured in mev . other parameters are taken as : @xmath101 mev , @xmath95 mev , @xmath100 mev , and @xmath151 . the inset in ( d ) depicts @xmath114 as a function of @xmath122 and @xmath110 . fig . 2 ( color online ) thermoelectric coefficients ( a ) @xmath128 , ( b ) @xmath112 , ( c ) @xmath113 , and ( d ) @xmath114 versus the dot s level position @xmath122 for different spin polarization @xmath121 in parallel configuration with @xmath152 mev . @xmath153 mev , other parameters are the same as in fig.1 . the inset ( 1 ) in ( d ) is an enlargement of the region near @xmath154 mev . the inset ( 2 ) depicts @xmath114 as a function of spin polarization @xmath121 with fixed @xmath155 mev . 3 ( color online ) thermoelectric coefficients ( a ) @xmath128 , ( b ) @xmath112 , ( c ) @xmath113 , and ( d ) @xmath114 versus the dot s level position @xmath122 in different configuration : antiparallel configuration and parallel configuration with @xmath151 . other parameters are the same as in fig . inset : enlargement of the region near @xmath154 mev . 4 ( color online ) thermoelectric coefficients ( a ) @xmath128 , ( b ) @xmath112 , ( c ) @xmath113 , and ( d ) @xmath114 versus the dot s level position @xmath122 for varied temperature @xmath156 in parallel configuration with @xmath134 mev and @xmath151 . the energy parameter @xmath157 is measured in mev . other parameters are taken as in fig . 2 . fig . 5 ( color online ) ( a ) electrical conductance @xmath128 and spin conductance @xmath158 , ( b ) thermopower @xmath113 and spin thermopower @xmath159 , ( c ) ( d ) charge figure of merit @xmath145 and spin figure of merit @xmath94 versus @xmath122 in parallel configuration with @xmath151 . the coulomb interactions @xmath95 mev in ( a ) ( b ) ( c ) and @xmath134 mev in ( d ) , other parameters are the same as in fig .	we study thermoelectric transport through double quantum dots system with spin - dependent interdot coupling and ferromagnetic electrodes by means of the non - equilibrium green function in the linear response regime . it is found that the thermoelectric coefficients are strongly dependent on the splitting of interdot coupling , the relative magnetic configurations and the spin polarization of leads . in particular , the thermoelectric efficiency can achieve considerable value in parallel configuration when the effective interdot coupling and tunnel coupling between qds and the leads for spin - down electrons are small . moreover , the thermoelectric efficiency increases with the intradot coulomb interactions increasing and can reach very high value at an appropriate temperature . in the presence of the magnetic field , the spin accumulation in leads strongly suppresses the thermoelectric efficiency and a pure spin thermopower can be obtained .
the dense plasma state is a common phase of matter in the universe and can be found in all types of stars @xcite and within giant planets @xcite . dense plasmas are commonly created during experiments involving high - power light sources such as , e.g. , the national ignition facility @xcite , and recently developed x - ray free - electron lasers ( xfels ) lcls @xcite and sacla @xcite . in dense plasmas , free electrons stay in the close vicinity of ions . the ions then can not any longer be treated as isolated species , as the screening by the dense environment shifts their atomic energy levels , leading to a reduction of the ionization potentials . this effect is known as ionization potential depression ( ipd ) . quantitative predictions of this effect are of crucial importance for a correct understanding and accurate modeling of any atomic processes occurring within a dense plasma environment @xcite , i.e. , for studies on warm dense matter @xcite , shock experiments @xcite , planetary science @xcite , inertial confinement fusion @xcite and studies of non - equilibrium plasmas created with xfels @xcite . several theoretical models have been developed to describe the ipd effect . an early development was the model proposed by ecker and krll ( ek ) @xcite for strongly coupled plasma , later extended to the weakly coupled regime by stewart and pyatt ( sp ) @xcite ( for more examples , see ref . however , until recently there were no experimental data available to verify the accuracy of these models whose predictions sometimes differed extensively . first experiments on the screening effect of plasma on atoms embedded in the plasma have been performed at lcls @xcite . xfels provide radiation of extremely high peak brightness and pulse duration shorter or comparable with the characteristic times of the electron and ion dynamics within irradiated systems . the dense electronic systems can quickly thermalize via electron electron collisions and impact ionization processes @xcite . because of the ultrashort pulse duration ( typically tens of femtoseconds ) , only a thermalized electron plasma is probed while the ions still remain cold . this provides access to the properties of a solid - density material at temperature of @xmath0@xmath1 k ( @xmath2100 ev ) . specifically , the experiments in refs . @xcite measured @xmath3-edge thresholds and @xmath4 emission from solid - density aluminum ( al ) plasma . they have been followed by another experiment at the high - power orion laser @xcite . this experiment investigated @xmath3-shell emissions from hot dense al plasma . both experimental teams tried to describe their findings with the ek and sp models . in the first experiment a disagreement of the measured @xmath3-edges with the extensively used sp model was claimed . a modified ek model was proposed to fit the experimental data @xcite . however , the data from the experiment on hot dense al plasma @xcite could only be described with the sp model . the ek model was found to be in clear disagreement with these data . this controversy shows a strong need for a rigorous and consistent theoretical approach able to calculate the ipd effect for plasmas in different coupling regimes . here we propose such an approach : a two - step hartree - fock - slater ( hfs ) method . this model derives the electronic structure of an ion embedded in the electron plasma from the finite - temperature approach @xcite , assuming thermalization of bound electrons within the free - electron plasma . it can also treat individual electronic configurations of plasma ions , which enables a description of discrete transitions . in this paper , we demonstrate that this model successfully describes laser - irradiated al solids under the conditions of the lcls @xcite and orion laser @xcite experiments . furthermore , we gain an improved understanding of the validity of the widely - used ek and sp models . in the first step , we apply the thermal hfs approach for a given finite temperature . here we assume that the electrons are fully thermalized . in general , xfel radiation induces non - equilibrium dynamics of electrons , for instance , in carbon - based materials exposed to hard x - rays @xcite . for the solid - density al plasma studied in the recent experiment @xcite , the incident photon energy is near the ionization threshold , ejecting electrons with low kinetic energy . due to the high density and low kinetic energy , electron cross sections are large , so that electrons equilibrate rapidly within the ultrashort pulse duration . for example , in the al plasma considered here , the energy of a photoelectron is less than 270 ev and the highest energy of an auger electron is about 1.4 kev . with a kinetic energy in this regime , the estimated time scale of electron thermalization is a few femtoseconds for diamond @xcite and is expected to be shorter for solid al due to higher impact ionization cross sections . therefore , we assume that the electrons are thermalized within the pulse duration of tens of femtoseconds . the standard hartree - fock and hartree - fock - slater ( hfs ) approaches @xcite treat electronic structure at zero temperature ( @xmath50 ev ) . they use the ritz variational principle for the ground - state energy . recently , thiele _ et al . _ @xcite proposed an extension of the standard hfs model , including plasma environment effects through the debye screening ( see also refs . however , this model is not applicable to plasmas with low temperature , where the debye screening approximation breaks down @xcite . also , it is intrinsically inconsistent to combine the @xmath6 ev approach with debye screening for a plasma with a non - zero electron temperature . to overcome this inconsistency , the electronic structure has to be derived from a finite - temperature approach . such a finite - temperature hartree - fock approach was proposed by @xcite . here we use the average - atom model @xcite , which is a variant of the finite - temperature approach . basically , it predicts average orbital properties and occupation numbers at a given temperature . there have been various implementations of the average - atom approach . depending on their treatment of the electronic structure of atoms , they can be categorized as quantum mechanical approaches , such as the hartree - fock - slater ( hfs ) method or local density approximation ( lda ) @xcite , or semi - classical approaches , such as the thomas - fermi ( tf ) method @xcite . there are also implementations with simplified super - configurations @xcite and with the screened hydrogenic model @xcite . the atomic potential within a plasma is usually based on the muffin - tin approximation @xcite or an extended model including ion ion correlation @xcite . these average - atom models have been applied to calculate physical quantities within plasmas , such as lowering of the ionization energy @xcite , photoabsorption processes @xcite , and scattering processes @xcite . our average - atom implementation presented here is based on the quantum mechanical approach with the muffin - tin approximation @xcite , but it benefits from a numerical grid technique , which will be discussed in detail later . the major distinction of our proposed method from previous average - atom models is not only the different numerical method . in this paper , we will develop a simple model to retrieve more complete information from the average - atom approach . we propose a two - step model : an average - atom calculation as the first step and a fixed - configuration calculation as the second step . from the average - atom approach , we obtain a grand - canonical ensemble at a given temperature with a simple electronic mean field associated with all possible configurations . using this information , we then calculate improved mean fields for selected configurations of interest . note that our two - step model is relatively inexpensive , in comparison with polyatomic density - functional calculations , and is easily applicable to any atomic species . this two - step model is described in detail in the three following subsections . schematic diagram of an atomic model in a solid or a plasma . the purple dotted curve is the isolated atomic potential , the green dashed curve is the crystalline potential , and the black thick curve is the muffin - tin - type potential . the muffin - tin flat potential is denoted as @xmath7 and the wigner - seitz radius is @xmath8 . ] to describe the electronic structure in a solid or a plasma , we employ a muffin - tin - type potential @xcite as depicted in fig . [ fig : potential_solid ] . influenced by the free electrons and neighboring ions , the atomic potential is lowered in comparison to that in an isolated atom . the sphere surrounding an atom is defined by the wigner - seitz radius , @xmath8 . if the solid consists of a single atomic species , then @xmath9 , where @xmath10 is the number density of ions in the solid . here we assume that the positions of the ions are fixed . therefore the wigner - seitz radius does not change during the calculation . we assume charge neutrality such that the ionic charge density outside the wigner - seitz sphere is compensated by the electron density . the net charge inside the wigner - seitz sphere is also zero on average , so that the potential outside is constant . we use this muffin - tin - type model for all our calculations . our implementation of the muffin - tin potential differs from the original model suggested by slater @xcite and its quantum - mechanical implementations @xcite . first , the constant potential outside the atomic sphere is self - consistently calculated in our model , whereas it is set to zero in other previous implementations @xcite . we refer to this constant potential as the muffin - tin flat potential , @xmath7 . second , we calculate both bound- and continuum - state wave functions with the same atomic potential , using numerical grids with a sufficiently large radius far from @xmath8 . this makes our method distinct from other implementations where a continuum state outside @xmath8 is usually given as a plane wave and special boundary conditions for both bound and continuum states are required . with the muffin - tin flat potential , we may consider different ionization processes in a solid or a cluster @xcite . the ionization energy is defined as the energy needed to transfer an electron to the continuum level located at @xmath11 , which corresponds to the binding energy measured with photoelectron spectroscopy . in a solid or a cluster , this process would be called _ outer - ionization_. on the other hand , there is already a continuum of states starting at @xmath7 , when the muffin - tin - type potential is imposed . it defines excitation into the continuum for @xmath12 , which would be called _ inner - ionization_. in metals like aluminum , the conduction band can be described by this continuum above @xmath7 and inner - ionization is a transfer of an electron bound to an atom ( narrow band ) into the conduction band . figure [ fig : potential_solid ] depicts schematically outer - ionization and inner - ionization processes for the muffin - tin - type potential . we solve the effective single - electron schrdinger equation with the muffin - tin - type potential ( atomic units are used unless specified otherwise ) , @xmath13 \psi(\mathbf{r } ) = \varepsilon \psi(\mathbf{r}),\ ] ] where the potential is the hartree - fock - slater ( hfs ) potential inside @xmath8 and is constant outside @xmath8 , @xmath14 } & \text{for } r \leq r_s , \\ { \displaystyle v(r_s ) } & \text{for } r > r_s , \end{cases}\ ] ] where @xmath15 is the nuclear charge , @xmath16 is the electronic density , and @xmath17 is the slater exchange potential , @xmath18 = - \frac{3}{2 } \left [ \frac{3}{\pi } \rho(\mathbf{r } ) \right]^{1/3}.\ ] ] we use a spherically symmetric electronic density , @xmath19 , so @xmath20 is also spherically symmetric . the potential for @xmath21 is given by the constant value of @xmath22 , fulfilling the continuity condition at the boundary @xmath23 . this constant potential defines the muffin - tin flat potential , @xmath24 . for an isolated atom without the plasma , we use the original hfs potential , which is eq . without applying the wigner - seitz radius and the muffin - tin flat potential , @xmath25 } .\ ] ] in contrast to plasmas and solids , the long - range potential in the isolated atom is governed by the coulomb potential [ @xmath26 where @xmath27 is the charge of the atomic system . however , it is well known that the original hfs potential does not have this proper asymptotic behavior due to the self - interaction term @xcite . to obtain the proper long - range potential for both occupied and unoccupied orbitals , we apply the latter tail correction @xcite . thus , the unscreened hfs potential for an isolated atom is given by eq . and replaced by @xmath28 only when @xmath29 . the orbital wave function @xmath30 is expressed in terms of the product of a radial wave function and a spherical harmonic , @xmath31 where @xmath32 , @xmath33 , and @xmath34 are the principal quantum number , the orbital angular momentum quantum number , and the associated projection quantum number , respectively . the radial wave function @xmath35 is solved by the generalized pseudospectral method @xcite . plugging eq . into eq . , we obtain the radial schrdingier equation for a given @xmath33 , @xmath36 u_{nl}(r ) = \varepsilon_{nl } u_{nl}(r).\ ] ] the hamiltonian and radial wave function in eq . are discretized on a nonuniform grid for @xmath37 . note that the boundary conditions are @xmath38 , and no additional boundary condition at @xmath8 is imposed . after diagonalizing the discretized hamiltonian matrix , one obtains not only bound states ( @xmath39 ) but also a discretized _ pseudocontinuum _ ( @xmath40 ) . with a sufficiently large radius @xmath41 , the distribution of these pseudocontinuum states becomes dense enough to imitate continuum states @xcite . plots of bound- and continuum - state energies of al at zero temperature , obtained by diagonalization of the discretized hamiltonian matrix . some of bound- and continuum - state wave functions are also plotted . the maximum radius is @xmath42 a.u . used in the calculation . ] figure [ fig : al_pot_wf ] shows some pseudocontinuum states as well as bound states ( @xmath43 and @xmath44 ) of al at zero temperature . the dense horizontal lines above @xmath7 are all pseudocontinuum states . here we use @xmath45 a.u . , which is much larger than @xmath46 a.u . from the al solid density ( 2.7 g/@xmath47 ) the number of grid points for @xmath48 is 200 and the number of partial waves is 31 ( @xmath49 . with these computational parameters , we obtain 6200 radial eigenstates . in the metallic al case , only 3 states ( @xmath50 , @xmath43 , and @xmath44 ) are bound states . all other eigenstates constitute a pseudocontinuum . we keep these computational parameters for all calculations throughout the paper . the numerical grid technique we use here attains advantageous simplicity in continuum - state calculations . one can transform the integration over positive energy into a summation over discrete states . there is no separation of the inside and outside regions , and therefore , no boundary condition at the wigner - seitz radius is needed . in contrast , other implementations of the muffin - tin model involve special boundary conditions at @xmath8 . for example , @xcite used the condition that the wave functions in the inner sphere are continuously connected to those in the outside region , and @xcite enforced the derivative of the wave function to vanish at the wigner - seitz radius . note that different boundary conditions at @xmath8 lead to different electronic structures as pointed out in ref . @xcite . the first step of our two - step hfs approach is an average - atom model calculation with the muffin - tin - type potential in eq . . we treat the electronic system using a grand - canonical ensemble at a finite temperature @xmath51 ( in units of energy ) . the electronic density @xmath52 is then constructed by @xmath53 where @xmath54 indicates the one - particle state index , i.e. , @xmath55 where @xmath56 is the spin quantum number , and @xmath54 runs over all bound and continuum states . here @xmath57 are fractional occupation numbers according to the fermi - dirac distribution with a chemical potential @xmath58 , @xmath59 where @xmath60 is the orbital energy for a given spin - orbital @xmath54 . the average number of electrons , @xmath61 , within the wigner - seitz sphere , @xmath62 is fixed to @xmath63 to ensure charge neutrality . this condition serves as a constraint to determine the chemical potential at the given temperature @xcite . in order to determine @xmath58 , one must find the root of the following equation , @xmath64 with @xmath58 obtained from eq . , @xmath52 is constructed from eq . . with @xmath52 , the updated atomic potential , as well as @xmath7 , is obtained from eq . . then , orbitals @xmath65 and orbital energies @xmath66 are calculated , using the new potential . again , a new @xmath58 is obtained from eq . . this self - consistent field ( scf ) procedure is performed until the results converge . note that there are only three input parameters in the calculation : element species ( @xmath15 ) , temperature ( @xmath51 ) , and solid density via the wigner - seitz radius ( @xmath8 ) . all other quantities such as orbitals , orbital energies , @xmath67 , @xmath58 , and @xmath7 are determined self - consistently . regarding the exchange potential at a finite temperature , various implementations have been proposed @xcite , but no unanimous expression has been identified . @xcite proposed a parameterization of the thermal exchange potential based on the local density approximation ( lda ) . @xcite proposed an interpolation between the zero - temperature slater potential and the high temperature limit . in the present calculations , we use the same potential as used in the zero - temperature calculation given by eq . . note that our approach can be easily combined with any type of exchange potential . we will discuss the dependence on different thermal exchange potentials in sec . [ sec : results ] . the second step in our two - step hfs approach is a fixed configuration calculation for bound electrons in the presence of the free - electron density . within the average - atom model , one can not obtain orbital energies of individual electronic configurations associated with different charge states . instead , orbital energies in the average - atom model represent averaged quantities for an averaged configuration with fractional occupational numbers . however , in a fluorescence experiment , for instance , one can see discrete transition lines corresponding to individual charge states @xcite , which are not accessible within the average - atom model . in order to describe individual electronic configurations within a plasma environment , we propose a fixed - configuration scheme . with the grand - canonical ensemble , one can calculate the probability distribution @xcite of all possible bound - state configurations ( see appendix [ appendix1 ] for details ) , @xmath68 } = \prod_{b}^\text{bound } \frac { e^ { - ( \varepsilon_{b } - \mu ) n_b / t } } { 1 + e^ { - ( \varepsilon_{b } - \mu ) / t } } , \ ] ] where @xmath69 = ( n_1 , \cdots , n_b ) $ ] indicates the fixed bound - state configuration , and @xmath70 is the number of bound one - electron states . here @xmath71 runs over all bound states @xmath72 and @xmath73 is an integer occupation number ( 0 or 1 ) in the bound - electron configuration . the probability of finding the charge state @xmath27 is given by the sum of all associated bound - state configurations , @xmath74}^{q } p_{[n_b]}.\ ] ] here @xmath69 $ ] runs over all possible bound - state configurations satisfying @xmath75 . from the probability distributions in eqs . and , one can choose one bound - electron configuration from the grand - canonical ensemble and perform a single scf calculation with this fixed configuration . for example , it is possible to choose the most probable configuration associated with the most probable charge state . the @xmath4 transition energy calculated from this configuration gives one discrete line in the x - ray emission spectrum . different configurations contribute to different transition lines , so measurement of these lines maps out the distribution of all possible configurations and charge states . here we focus on individual bound - electron configurations . even though the bound - state electronic structure is influenced by the presence of the plasma electrons , we assume that it is not sensitive to detailed free - electron configurations in the plasma . therefore , once we choose one bound - electron configuration , we calculate the free - electron density as an average of all possible free - electron configurations for the given bound - electron configuration ( see appendix [ appendix2 ] ) , @xmath76 which turns out to be independent of the bound - electron configuration selected . this free - electron density is self - consistently obtained in the first step and is kept fixed in the second step . the bound - electron density is constructed with a fixed electron configuration , @xmath77 then , the total electron density is constructed as the sum of the bound and free - electron densities , @xmath78 with this total electron density , we perform a hfs calculation using a microcanonical ensemble . in this case , @xmath79 is self - consistently updated , whereas @xmath80 remains fixed during the scf procedure . this approach allows the bound electrons in a given configuration to adjust to the presence of the plasma electrons . in this section , we apply the two - step hfs model to laser - irradiated al solid @xcite . the two - step procedure is carried out as follows . first , we perform a finite - temperature hfs calculation to determine the temperature needed to achieve a certain average charge state within the plasma . we then use the free - electron density @xmath81 obtained from the finite - temperature calculation to perform a fixed - configuration hfs calculation for the given charge state to determine the energy of the @xmath50 orbital and the energy of the energetically lowest @xmath54 orbital above @xmath44 ( always referred to as @xmath82 , whether it is bound or not ) . both the first and second steps of our two - step hfs model are implemented as an extension of the xatom toolkit @xcite , which can calculate any atomic element and any electronic configuration within a non - relativistic framework . all calculations are performed with the computational parameters stated in sec . [ sec : first - step ] and are fully converged . .[table : al]electronic structure of al metal and al plasma . all energies in our present calculations are subtracted by @xmath7 , in order to compare with other theoretical data . al solid density is @xmath83 g/@xmath47 and @xmath84 is the average charge state . energies are in ev . [ cols="^ , > , > , > , > " , ] as shown in the previous subsection , we determine from the first step for a given temperature : a ) probabilities of all individual electronic configurations associated with different charge states , and b ) the free - electron density . for the lcls conditions ( @xmath8580 ev ) corresponding to the strongly and moderately coupled plasma regimes , the average charge state , the most probable charge state , and the most probable configuration of this charge state are listed in table [ table : config ] . energy levels as a function of the charge state for aluminum . ( a ) the @xmath82 orbital energy and the threshold energy to the continuum ( @xmath7 ) calculated with the two - step hfs approach . the threshold energy to the continuum for isolated ions ( @xmath11 ) is shown for reference . ( b ) the @xmath50 orbital energies without the plasma ( unscreened hfs calculations ) and with the plasma ( two - step hfs calculations ) . ] figure [ fig : al_energy ] shows ( a ) the resulting @xmath82 ( or the lowest - energy @xmath54 state in the continuum ) orbital energies and the muffin - tin flat potential calculated by the two - step hfs scheme with the free - electron density , and ( b ) the @xmath50 orbital energies with and without the plasma , as a function of the charge state . all those energies are lowered as the charge state increases . note that the @xmath82 energy lies right at the threshold to the continuum , i.e. , it is not bound to a single atom . the only exception is al@xmath86 , where @xmath82 lies @xmath873.4 ev below the threshold . for an isolated atom or ion , calculated in the unscreened hfs approach , the threshold energy to the continuum is constant ( @xmath11 ) for all charge states . for a solid , the threshold energy to the continuum ( @xmath7 ) decreases by 5 ev from al iv to al viii . lowering of the @xmath50 binding energies due to the plasma environment ( 44107 ev ) is much larger than the lowering of the threshold energy ( 1116 ev ) . for @xmath88 ev , the difference between the threshold energy to the continuum in fig . [ fig : al_energy](a ) and the @xmath50 orbital energy in fig . [ fig : al_energy](b ) gives the @xmath3-shell ionization potential . in our approach , both the @xmath50 orbital energy and the threshold energy are modified by the plasma environment . @xmath3-shell thresholds for aluminum as a function of the charge state calculated with the two - step hfs model . the experimental data are taken from ref . unscreened hfs refers to calculations for isolated ions . ] in the lcls experiment on al plasma @xcite , @xmath4 fluorescence was detected and spectrally resolved as a function of the incoming photon energy . in this way , the incident - photon - energy threshold for the formation of a @xmath3-shell hole was determined for each energetically resolvable charge state . a @xmath3-shell hole can be created for @xmath88 ev by inner - ionization or photo - excitation into the @xmath82 orbital if @xmath82 is bound . figure [ fig : fixed_config ] shows the calculated @xmath3-shell ionization thresholds ( photo - excitation for al@xmath86 ) , in comparison with the experimental results @xcite . we also plot the @xmath3-shell ionization thresholds for the unscreened hfs method ( isolated ions ) and the average - atom model . for al@xmath86 , the resonant excitation into @xmath82 is below the ionization threshold by just @xmath873.4 ev , which may not be resolvable due to the lcls energy bandwidth of @xmath877 ev in experiment @xcite . as shown in fig . [ fig : fixed_config ] , the two - step hfs calculation yields good agreement with the experimental data . however , the average - atom model alone fails in reproducing experiment , especially for high charge states . in experiment , each discrete fluorescence line selects only one charge state and the @xmath3-shell threshold is assigned to this specific charge state . the fixed - configuration scheme in our two - step model properly describes this selection of the @xmath3-shell threshold , whereas the average - atom model with the configuration averaging does not . all calculated energies were shifted by + 21.5 ev , according to the difference between the inner - ionization energy calculated at @xmath6 ev ( 1538.1 ev ) and the experimental binding energy ( 1559.6 ev ) @xcite . this constant energy shift is a model assumption for comparing our results to the experimental data . note that the absolute accuracy of hfs binding energies is typically about 1% . clearly , in order to improve the description , one would require a treatment of the electronic structure beyond the mean - field level . however , it may be anticipated that such an approach would be much less efficient than the present hfs theory . on the other hand , the error bar in the two - step hfs model in fig . [ fig : fixed_config ] indicates variation from different thermal exchange potentials used in our calculations . we have tested the thermal exchange potentials of @xcite and @xcite , following the same two - step procedure as described in sec . [ sec : theory ] , and found that the maximum deviation from the results with the zero - temperature potential is about 12 ev . plots of the ionization potential depression . ( a ) comparison for all orbitals of al calculated with the two - step hfs approach . ( b ) comparison of various methods : two - step hfs model , modified ek model , and sp model . the predictions of the latter two are taken from ref . @xcite . ] by taking the difference of the ionization potentials with and without the plasma environment , we can examine the lowering of the ionization potentials not only for @xmath3-shell electrons but also for electrons in other subshells . for individual bound orbitals of al ions , we obtain the ipd shown in fig . [ fig : ipd](a ) . for isolated atoms , the ionization potential is given by @xmath89 where @xmath90 is the @xmath91th orbital energy from the unscreened hfs calculation . for atoms in the plasma for @xmath88 ev , the inner - ionization potential is calculated by @xmath92 , where both @xmath7 and @xmath93 are obtained from the two - step hfs calculation . since the plasma screening affects each orbital differently , it is expected that ipds for individual orbitals are different . our results show that the ipd for @xmath50 is higher than the ipd of @xmath43 and @xmath44 by @xmath873 ev , but there is almost no difference in the ipds for orbitals with the same principal quantum number ( @xmath43 and @xmath44 ) . this trend is similar to that observed in ref . @xcite for the debye - screened hfs model . figure [ fig : ipd](b ) depicts a comparison of various theoretical ipd models . the results of the stewart - pyatt ( sp ) model and the modified ecker - krll ( ek ) model are taken from ref . the original ek model @xcite and the sp model @xcite for lowering of the ionization energy have been widely used in the past decades and are implemented in several codes , e.g. , flychk @xcite or lasnex - dca @xcite . ecker and krll @xcite have described lowering of the ionization potential as being due to the presence of an electric microfield . in their model , there is no difference among the ipds of individual orbitals , and the ionization potential is considered as the difference between the ground - state energy of the charge state @xmath27 and that of the charge state @xmath94 , which corresponds to the outermost - shell ionization potential in our calculations ( @xmath44 for the al plasma ) . a modified version of the ek model ( mek ) has been proposed in refs . @xcite by employing an empirical constant to fit the experimental data . figure [ fig : ipd](b ) shows that neither the mek model nor the sp model are close to our two - step hfs approach . @xmath4 transition energies of aluminum ions obtained with the two - step hfs model . they are compared to the experimental data taken from ref . the gray areas indicate the full - width - at - half - maximum from experiment . ] in figure [ fig : transition_energy ] , we show the peak positions of the @xmath4 fluorescence lines for al iv up to al viii . the experimental data are taken from ref . the transition energies are calculated from the differences of the @xmath44 and @xmath50 orbital energies in the fixed - configuration scheme . the fixed bound - electron configuration associated with a given charge state is chosen from table [ table : config ] . all calculated energies are shifted by + 21.5 ev , according to the difference between the al iv transition energy calculated with the average - atom model at @xmath6 ev ( @xmath95 ev ) and the experimental transition line ( @xmath96 ev ) @xcite . the error bar in the two - step hfs model shows variation from usage of different thermal exchange potentials , and the maximum deviation is about 2 ev . our results show small deviations ( @xmath97 ev ) from the experimental transition energies . here we show transition energies for only one configuration associated with a given charge state . however , different configurations , as listed in table [ table : config_probability ] at @xmath98 ev , would give rise to different transition lines . for instance , the ground configuration ( @xmath99 ) of al viii at @xmath98 ev gives a fluorescence line of + 3 ev higher in energy than the most probable configuration ( @xmath100 ) . the muffin - tin flat potential ( @xmath7 ) and @xmath82 orbital energy ( @xmath101 ) of h - like and he - like al as a function of the solid density at @xmath102 ev . ] our two - step scheme is applicable not only in the strongly coupled plasma regime but also in the weakly coupled plasma regime . recently , @xcite used the high - power orion laser to create compressed al plasmas with high temperature and measured ly@xmath103 and he@xmath103 lines to diagnose the al plasmas created . in fig . [ fig : orion ] , the two - step model calculations show that the @xmath82 state of compressed al merges with the continuum ( @xmath40 ) as the al density increases . the electronic temperature is 700 ev , close to the orion experimental condition @xcite . for the bound - electron part in the second step , the h - like al is calculated with the exact one - electron potential and the he - like al is calculated with the exact hartree - fock potential . for such a high temperature , the plasma electron density contributes to only the direct coulomb interaction @xcite . when the solid density @xmath104 is greater than @xmath8712 g/@xmath47 , the @xmath82 state of h - like al is no longer bound , so the ly@xmath103 line would disappear . likewise , the he@xmath103 line would disappear after @xmath105 g/@xmath47 . our results are consistent with the experimental finding of no @xmath106 transitions occurring at @xmath10710 g/@xmath47 . when we use different thermal exchange potentials @xcite , the merging point of @xmath82 becomes @xmath878 g/@xmath47 and @xmath8711 g/@xmath47 for he@xmath103 and ly@xmath103 , respectively . the sp model predicts delocalization of @xmath106 levels for @xmath108 g/@xmath47 , whereas the ek model prediction was found to be in clear disagreement with the experimental data @xcite . in this work , we have extended the standard hfs approach for calculating atomic energy levels for ions embedded in a plasma , taking into account plasma screening . our two - step hfs model includes : ( i ) the average - atom model to obtain the free - electron density at a given temperature and ( ii ) the fixed - configuration model taking into account the free - electron density . our current analysis focused on al plasmas created by lcls @xcite and orion laser @xcite , covering both strongly and weakly coupled plasma regimes . our two - step hfs results on the @xmath3-shell threshold energies of different charge states within al plasma are in good agreement with the lcls experimental data @xcite . references @xcite measured fluorescence threshold energies that were then used to extract ipds by combining the data with a specific theory model describing the unscreened ionization potentials . thus , the estimated ipds relied on the theory of the unscreened case , which hinders direct comparison with ipd models . in contrast , our model computes the energy shifts of all individual orbitals with and without plasma screening , thus providing ipds in an internally consistent manner . our calculated valence ipds lie between the sp and mek models . hence , we can not confirm that the performance of the mek model is superior to that of the sp model , as suggested in ref . moreover , in the high temperature regime , our prediction for the @xmath82 state is in good agreement with the sp model and reproduces orion experimental data @xcite . these results show that , with our proposed two - step hfs approach , a reliable and relatively inexpensive calculation of atomic properties within plasmas can be performed both for weakly and strongly coupled plasmas . we therefore expect that our model will be a useful tool for describing new data from plasma spectroscopy experiments . we thank hyun - kyung chung , orlando ciricosta , nikita medvedev , sam m. vinko , and justin s. wark for helpful discussions . during the review process , a new preprint treating the same ipd problem appeared ( b. j. b. crowley , arxiv:1309.1456 ) . the partition function of the grand - canonical ensemble for fermions is given by @xmath109 where @xmath110 is the inverse of the temperature , @xmath111 , and @xmath112 is the occupation number of the @xmath54th one - electron state . here @xmath54 includes all bound and continuum one - electron states , and the summation runs over all possible configurations @xmath113 . for fermions , @xmath112 is either 0 or 1 . then , the probability of finding one specific configuration @xmath114 is @xmath115 let us consider a fixed configuration , more precisely , a fixed bound - electron configuration @xmath69 = ( n_1 , \cdots , n_b ) $ ] , where @xmath70 is the number of the bound states . a general fock - space configuration consistent with the fixed bound - electron configuration is @xmath116 here @xmath73 for @xmath117 is a fixed occupation number , while @xmath118 is either 0 or 1 for @xmath119 . the probability of finding the fixed configuration @xmath69 $ ] is calculated by summing over all such configurations @xmath120 , @xmath121 } & = \sum_{\ { n_p \ } = \ { n_b ; n_{p ' } \ } } p_{\lbrace n_p \rbrace } \nonumber \\ & = \frac{1}{y } \prod_{b=1}^b e^{-\beta(\varepsilon_b - \mu ) n_b } \sum _ { \ { n_{p ' } \ } } ^\text{all } \prod_{p'=b+1}^{\infty } e^{-\beta(\varepsilon_{p ' } - \mu ) n_{p ' } } \nonumber \\ & = \frac { \prod_{b=1}^b e^{-\beta(\varepsilon_b - \mu ) n_b } \prod_{p'=b+1}^{\infty } \left ( 1 + e^{-\beta(\varepsilon_{p ' } - \mu ) } \right ) } { \prod_{b=1}^b \left ( 1 + e^{-\beta(\varepsilon_b - \mu ) } \right ) \prod_{p'=b+1}^{\infty } \left ( 1 + e^{-\beta(\varepsilon_{p ' } - \mu ) } \right ) } . \nonumber\end{aligned}\ ] ] in the above calculation , since the occupation numbers in the bound states are fixed , they can be factored out and the remaining parts in the numerator and the denominator are canceled out . thus , the probability reduces to @xmath121 } & = \prod_{b=1}^{b}\frac { e^{-\beta(\varepsilon_b-\mu ) n_b}}{1 + e^{-\beta(\varepsilon_b-\mu)}},\end{aligned}\ ] ] which corresponds to eq . . this is for spin - orbitals @xmath71 , i.e. , @xmath73 is either 0 or 1 . for practical purposes , we express the probability in terms of subshells , @xmath122 } = \frac { \prod_{i=1}^{n_b } \binom { 4 l_i + 2 } { n_i } e^ { -\beta ( \varepsilon_{i } - \mu ) n_i } } { \prod_{i=1}^{n_b } \left ( 1 + e^ { -\beta ( \varepsilon_{i } - \mu ) } \right)^{(4 l_i + 2 ) } } , \ ] ] where @xmath123 is the subshell index , @xmath124 is the number of bound subshells , and @xmath125 = ( n_1 , \cdots , n_{n_b } ) $ ] . for the @xmath123th subshell , @xmath126 is the orbital energy , @xmath127 is the orbital angular momentum quantum number , and @xmath10 is the occupation number ( @xmath128 ) . the probability of finding the charge state @xmath27 is given by summing over all possible configurations associated with @xmath27 , @xmath129}^q p_{[n_b ] } = \sum_{[n_b]}^q \prod_{b=1}^{b } \frac { e^{-\beta(\varepsilon_b-\mu ) n_b}}{1 + e^{-\beta(\varepsilon_b-\mu)}},\ ] ] where @xmath69 $ ] satisfies @xmath130 . this is eq . . it is straightforward to verify @xmath131 . the average charge state @xmath84 is given by @xmath132}^q \left ( z - \sum_{p=1}^{b } n_p \right ) \prod_{b=1}^{b } \frac { e^{-\beta(\varepsilon_b-\mu ) n_b } } { 1 + e^{-\beta(\varepsilon_b-\mu ) } } \nonumber \\ & = z - \sum_{b=1}^b \tilde{n}_b(\mu , t ) , \ ] ] which is used to calculate @xmath84 in tables [ table : al ] and [ table : config ] . here we calculate the total electron density for a fixed bound - electron configuration , @xmath133 = ( n_1 , \cdots , n_b ) $ ] . the total density is chosen by averaging densities over all configurations @xmath120 that have @xmath133 $ ] in common : @xmath134 } = \sum _ { \ { n_p \ } = \ { n_b ; n_{p ' } \ } } \rho_{\ { n_p \ } } w _ { \ { n_p \ } } , \ ] ] where @xmath135 is a statistical weight , @xmath136 the total density for @xmath113 decomposes into bound - electron and free - electron densities , @xmath137 plugging eqs . and into eq . , we evaluate the total density for @xmath69 $ ] . the bound - electron density for @xmath69 $ ] is then @xmath138 } & = \left ( \sum_{b=1}^b \left\| \psi_b(\mathbf{r } ) \right\|^2 n_b \right ) \sum _ { \ { n_{p ' } \ } } ^\text{all } \prod_{p'=b+1}^{\infty } \frac { e^{- \beta ( \varepsilon_{p ' } - \mu ) n_{p ' } } } { 1 + e^{- \beta ( \varepsilon_{p ' } - \mu ) } } \nonumber \\ & = \sum_{b=1}^b \left\| \psi_b(\mathbf{r } ) \right\|^2 n_b,\end{aligned}\ ] ] and the free - electron density for @xmath69 $ ] is given by @xmath139 } & = \sum _ { \ { n_{p ' } \ } } ^\text{all } \!\ ! \left ( \sum_{p'=b+1}^{\infty } \left\| \psi_{p'}(\mathbf{r } ) \right\|^2 n_{p ' } \!\ ! \right ) \!\ ! \prod_{p'=b+1}^{\infty } \frac { e^{- \beta ( \varepsilon_{p ' } - \mu ) n_{p ' } } } { 1 + e^{- \beta ( \varepsilon_{p ' } - \mu ) } } \nonumber \\ & = \sum_{p = b+1}^{\infty } \left\| \psi_{p}(\mathbf{r } ) \right\|^2 \frac { 1 } { e^{\beta ( \varepsilon_{p } - \mu ) } + 1}.\end{aligned}\ ] ] therefore , the total electron density for configuration @xmath69 $ ] is determined from the average - atom calculation ( the first step of our two - step approach ) with the grand - canonical ensemble , @xmath140 } & = \rho_{b,[n_b ] } + \rho_{f,[n_b ] } \\ & = \sum_{b=1}^b \left\| \psi_b(\mathbf{r } ) \right\|^2 n_b + \sum_{p = b+1}^{\infty } \left\| \psi_{p}(\mathbf{r } ) \right\|^2 \tilde{n}_p(\mu , t).\end{aligned}\ ] ] in the second step of our two - step approach , we separate out the free - electron density , @xmath141}$ ] , which yields eq . . 77ifxundefined [ 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 @noop _ _ , ( , , ) link:\doibase 10.1088/0741 - 3335/51/12/124014 [ * * , ( ) ] link:\doibase 10.1088/0004 - 637x/726/1/15 [ * * , ( ) ] link:\doibase 10.1063/1.3116505 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1063/1.3078101 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1038/nature10746 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop _ _ ( , , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop _ _ , , vol . ( , , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) in @noop _ _ , , vol . , ( , ) @noop * * , ( ) @noop , cfel , desy , hamburg , germany , , @noop * * , ( ) @noop _ _ ( , , ) http://xdb.lbl.gov/ [ `` , '' ] ( ) @noop * * , ( ) link:\doibase 10.1016/0022 - 4073(87)90039 - 2 [ * * , ( ) ]	the charged environment within a dense plasma leads to the phenomenon of ionization potential depression ( ipd ) for ions embedded in the plasma . accurate predictions of the ipd effect are of crucial importance for modeling atomic processes occurring within dense plasmas . several theoretical models have been developed to describe the ipd effect , with frequently discrepant predictions . only recently , first experiments on ipd in al plasma have been performed with an x - ray free - electron laser ( xfel ) , where their results were found to be in disagreement with the widely - used ipd model by stewart and pyatt . another experiment on al , at the orion laser , showed disagreement with the model by ecker and krll . this controversy shows a strong need for a rigorous and consistent theoretical approach to calculate the ipd effect . here we propose such an approach : a two - step hartree - fock - slater model . with this parameter - free model we can accurately and efficiently describe the experimental al data and validate the accuracy of standard ipd models . our model can be a useful tool for calculating atomic properties within dense plasmas with wide - ranging applications to studies on warm dense matter , shock experiments , planetary science , inertial confinement fusion and studies of non - equilibrium plasmas created with xfels .
the astronomer s typical role is that of a distant observer , trying to reconstruct an object s physical properties from afar . the key to inferring such properties is knowledge of the object s distance . after all , bright objects can look dim when seen from afar , and large objects can look small . only when an object s distance to us is known can we relate observational data to physical properties : the object s angular diameter to its physical size , or the amount of light we receive from an object to its intrinsic brightness . unfortunately , estimating cosmic distances is a difficult task . as no single method can cover the whole gamut of astronomical distances , astronomers have constructed a _ cosmic distance ladder _ that proceeds step by step , method by method , from distances within our solar system to the largest cosmic distances of billions of light - years @xcite . the fundamental concepts of this distance ladder are part of the curriculum of introductory astronomy courses @xcite and commonly considered basic astronomical knowledge essential for understanding how astronomers know what they know . beyond the realm of nearby stars ( with distances up to some tens of thousands of light - years ) , astronomical distance measurement relies on _ standard candles _ , as follows . assume that a cosmic object emits the total energy @xmath0 per second . @xmath0 is called the object s _ luminosity_. furthermore assume that the emission is isotropic : spread evenly over all directions , with no direction favoured or disfavoured . under such conditions , it is readily shown that the apparent brightness of the object , as seen by an observer at distance @xmath1 , follows an inverse - square law . let us consider the energy emission of the object as a serious of bursts , each of ( very brief ) duration @xmath2 , seemlessly following each other . the energy @xmath3 emitted in one such burst will be spread evenly on an expanding spherical shell . as it reaches the observer , it will have spread evenly over a spherical surface of radius @xmath1 , that is , with area @xmath4 . hence , the fraction of energy @xmath5 that falls into a detector pointed directly at the cosmic object by the observer during the time @xmath2 will be proportional to the ratio of the detector s area @xmath6 and the spherical shell area , @xmath7 astronomers typically quote the energy flux @xmath8 , that is , the energy detected per unit detector area per unit time . in our case , this means dividing the energy @xmath5 detected by the detection time and detector area , leading to the simple formula @xmath9 linking flux , luminosity and distance . flux is , in principle , a measurable quantity . a _ standard candle _ is defined as an object whose luminosity @xmath0 is known , or can be deduced from observable quantities . for such a standard candle , the measurable quantity @xmath8 and the known quantity @xmath0 can be used to solve equation ( [ inversesquare ] ) for the distance @xmath1 : the apparent brightness ( represented by the flux ) and the known intrinsic brightness @xmath0 can be used to determine the object s distance . various standard candles are used in contemporary astronomy , using diverse physical laws or heuristic relationships to estimate luminosities . the first astronomical standard candles were the cepheids : variable stars whose period of variation ( readily measured ) is correlated with their luminosity . first hints of this correlation were discovered by henrietta swan leavitt in 1908 , and a more precise formulation given by leavitt in 1912 . with the first calibration by ejnar hertzsprung in 1913 , cepheids became a distance indicator and , within the following years , they became instrumental in determining the first extragalactic distance scales , and in demonstrating that our galaxy is just one among many @xcite . modern astronomy uses a variety of standard candles . _ primary distance indicators _ are those that can be calibrated within our home galaxy . they include cepheids , but also other types of variable stars with period - luminosity relations : rr lyrae and mira type variables . other standard candles make use of luminosity - color diagrams ( equivalently : luminosity - temperature diagrams ) of star clusters : both for main sequence stars in the hertzsprung - russell diagram and for so - called red clump stars , this allows to deduce stellar luminosities within a fairly narrow range . eclipsing binaries allow for another kind of estimate : a doppler measurement of orbital velocities , combined with the light - curve data , leads to an estimate of the primary star s radius ; this , when combined with the star s temperature ( deduced from the spectrum ) and the stefan - boltzmann law leads to an estimate of the star s luminosity@xcite . _ secondary distance indicators _ can be observed out to much greater distances , but need to be calibrated using distant galaxies for which primary distance indicators are known . a number of secondary standard candles use overall properties of galaxies . examples are the tully - fisher relation , a heuristic correlation between a spiral galaxy s rotation speed ( via measured via spectral line width ) and stellar luminosity , the faber - jackson relation correlating elliptical galaxies random stellar velocities ( measured again via spectral line widths ) with total luminosity , and the fundamental plane relation , a refinement of the faber - jackson relation that takes into account surface brightness as an additional observable parameter . the method of surface brightness fluctuations , on the other hand , makes use of the fact that a galaxy s stellar luminosity is made up of the light of myriads of individual stars ; even at distances where individual stars can no longer be identified , this gives galaxies a slightly grainy appearance that varies with the distance between galaxy and observer @xcite . the most famous secondary distance indicators are supernovae of type ia ( sn ia ) : thermonuclear explosions of white dwarf stars , for which there is a correlation between the time scale of the explosion ( which can be read off of the supernova light curve ) and its peak luminosity . studies of the distance - redshift relation using these supernovae led to the discovery that the expansion of the universe is accelerating ; a discovery that was honoured with the nobel prize in physics in 2011 @xcite . typical classroom demonstrations of the inverse square law involve a photodetector , such as a small solar cell , placed at varying distances from a light source @xcite . here , we describe a different kind of hands - on activity , which treats street lights as standard candles . the activity uses a digital camera and freely available software . it is suitable as a high school class project as part of a physics course , or as an astronomy club activity . the exercise works best along a long , straight road ; for our worked - out example , we have chosen a straight road in heidelberg , germany that measures 660 m end to end . in the following , we will determine reference distances between the observer ( represented in this case by an slr camera ) and the target objects ( the street lights ) using the distance measuring tool in google maps . this would be akin to testing a particular type of standard candle against another , well - established distance indicator . observations of street lights will usually be made from a lower height than that of the street light s lamp itself . for that reason , it is important that the street lights be of a type where the role of perspective is negligible . this rules out street lights whose light shines out downward from a horizontal surface ; such a surface would present a larger section to an observer closer to and below the light than to a more distant observer , representing a systematic change in luminosity detrimental to standard candle measurements . spherical street lights would be ideal ; the street lights used here were of a truncated - cone type shown in fig . [ streetlight ] , which are fairly robust with respect to such changes of perspective . using a camera offset against the line of street lights at one end of the street , the line of lights can be captured in a single image , parts of which are shown in figure [ photometricimage ] . in order to read off the different street lamps apparent brightness , pixel values within the image need to be linear if pixel a has received twice as much light as pixel b , pixel value a must be twice as large ; pixel value must be proportional to the amount of light received by each pixel . for a dslr camera , this is not a given . images that are saved in jpg format , for instance , are processed in a non - linear fashion that is tailored to human expectations for images showing everyday situations . the images are meant to look natural , but they are not intended for quantitative measurements . thus , the first challenge is to save the dslr image in a way that allows for faithful measurements of brightness ( in other words , for good photometry ) . the first step is to make the camera save the image in a raw format , which does not throw any of the brightness information away ; the second is to transform the image into a format readable by software which can make simple brightness measurements : marking an image region and reading out the sum , or equivalently the mean , of all the pixel brightness values within that region . for astronomical photometry , there is a standard method called _ aperture photometry _ , which is of relevance for this situation . it amounts to summing up pixel brightness values within a region of area @xmath6 containing the object of interest , giving a total brightness @xmath10 . this total brightness is due not only to light emitted by the object , but also to the background brightness of the sky contained within @xmath6 ; this background brightness is due in part to light scattered by the atmosphere , in part due to unresolved cosmic objects other than the object under scrutiny . the background contribution to @xmath10 can be estimated by measuring the average pixel value @xmath11 within a surrounding annulus outside the area @xmath6 . this makes @xmath12 a good estimate for the background brightness contribution to @xmath10 , and @xmath13 a good estimate for the brightness @xmath14 due to the object of interest . similarly , in the image showing the street lamps , it is necessary to estimate background brightness @xmath11 ; we do so by measuring mean pixel brightness in a rectangular area directly above each lamp ( hence shielded from the lamp s own light by the lamp shade ) . the resulting brightness values @xmath15 using equation ( [ aperturephotometry ] ) are uncalibrated . as a simple way of calibration , we took the distance @xmath16 between the camera and the first street light , as measured using google maps , as given ; by the inverse square law ( [ inversesquare ] ) , the distance of the @xmath17th street light should then be @xmath18 the upper plot in figure [ firstdistanceplot ] shows the distances inferred by treating the street lights as standard candles , using equation ( [ streetlightsc1 ] ) . both these distances @xmath19 and the google map distances @xmath20 are plotted against street light number . the lower plot shows the relative difference @xmath21/d_{gm}(i)$ ] between the two . ideally , the two distance values should be the same for each street light . yet the plot shows marked discrepancies for a few of the lights ( in particular numbers 11 , 12 , 16 and 17 ) . what has gone wrong ? using street lights as standard candles leads to problems similar to those involving astronomical standard candles in general ; in other words : from analysing problems with their street light measurements , pupils can learn about confounding factors that occur in analyses of astronomical standard candles , as well . ( [ streetlightsc1 ] ) assumes that all street lights have the same luminosity but what if there are slight ( or even not so slight ) variations ? in those cases , distance values calculated via that equation will be skewed systematically street lights that are intrinsically fainter will appear as further away than they really are ; abnormally bright street lights will appear as significantly closer . this is a key problem for astronomical standard candles : how homogeneous is a population of objects ? do the objects under consideration indeed all have the same luminosities , or are there variations and if there are , are there measurable properties correlated with these variations , which can be used to predict the luminosity values for each of the objects ? the most famous example of a standard candle that underwent this kind of correction are , once more supernovae of type ia : initially , they were thought to be standard candles with approximately constant luminosity . but work in the early 1990s showed that there is a correlation between the time scale of a supernova s light curve and its peak luminosity : light curves that decline fastest are characteristic of the dimmest supernova explosions @xcite . in the case of street lights , it will hardly be worth our while to formulate intricate physical models , but instead we have the advantage of being able to document brightness variations directly . the proper way of doing this would be to take an image of each street light from a standard distance and , to account for emission anisotropies , from the same direction as for the image used to measure relative street light brightnesses . instead , we have followed a simplified procedure , taking an image of each street light from an approximately standardised distance ( not measured precisely , but counted off in steps ) in the direction along the street ( which deviates from our viewing direction for the brightness measurements by @xmath22 for the first light , rapidly decreasing to less than @xmath23 by the time we have reached the 4th light ) . luminosities relative to the first light , @xmath24 for @xmath25 the luminosity of the @xmath17th street light , as estimated by that method , are shown in fig . [ relativeluminosities ] . taking into account such intrinsic variations of luminosity , the new formula linking distance and apparent brightness is @xmath26 the upper plot in fig . [ luminositycorrecteddistanceplot ] shows , for each street light , the standard candle distance calculated in this way as well as the google map distance ; the relative differences of the two are once more shown in the lower plot . this has taken care of the outliers 16 and 17 ; direct inspection of those two street lights shows that they have been shielded so as not to disturb inhabitants of the nearby houses which , at that point of the street , have windows rather close to those lights . what , then , is the remaining problem with lanterns 11 and 12 ? an image taken at greater magnification reveals another problem that has an analogue for astronomical standard candles : _ extinction _ , that is , the absorption / scattering of light as it passes through intervening matter . in astronomy , extinction is typically due to gas and dust that are located between the observer and the target object . astronomical extinction can be estimated making use of the fact that different wavelengths of light are affected by extinction in different ways . in our case , taking an image of the more distant street lights directly shows extinction due to shading by branches from overhanging trees , as can be seen in fig . [ treeshading ] for the street lights 11 , 12 , 13 and 15 . going by direct visual inspection of the image , those four lights are the only ones affected . the fraction of the total area shaded by the branches can be estimated in the following simple way : measure the mean pixel brightness @xmath27 for the whole street lamp , including the obscured parts . measure the mean pixel brightness for selected parts of the lamp that are not obscured , and use their average ( weighted with area ) to estimate the lamp s unobscured mean brightness @xmath28 . then @xmath29 yields a good estimate for the fraction of lamp light lost due to obscuration . values for the four affected lamps are given in table [ obscurationtable ] . cc 11 & 57.4% + 12 & 48.9% + 13 & 21.6% + 15 & 42.6% in our case , extinction means that we need to multiply the luminosity @xmath25 of the @xmath17th street light by @xmath30 in order to calculate the expected amount of light received . the modified standard candle equation reads @xmath31 the result , once more in comparison with the google map distances , is shown in fig . [ finalcorrecteddistances ] . for more than two thirds of the street lights , the agreement is now better than 10% ; in astronomy , agreement around the 10% mark is commonly considered a good ( or even great ) result . there is still a tendency of calculated inverse - square law distances to lie above the green line , hinting at systematic errors . the simplest such error would be a problem with the calibration which , after all , relies on a single google maps measurement ( namely that of the distance between the camera location and the first street light ) and on estimates for the street lanterns relative luminosities . the set - up described here is suitable for a basic in - the - field demonstration of inverse square law distance determinations . where higher precision is desired , a number of opportunities for improvements suggest themselves : instead of comparing the distances between street lights with an online map , we could directly determine those distances using tape measures . for the luminosities , we could ensure a more strictly standardised measurement by taking into account the exact perspective of the camera to suppress effects of direction - dependent lighting . for the extinction estimate , we could use images taken at an even greater focus length . street lights as standard candles provide a useful hands - on activity that demonstrates to the learners the basic principles of distance measurement via standard candles . in contrast with laboratory measurements where one varies the distance between one light source and the observer / detector , the activity can also serve to illustrate some of the main potential problems with standard candle observations : unforeseen variations of standard candle luminosity and extinction effects . the tools used , namely a dslr and simple image processing software , are similar to their professional astronomical counterparts as to make the activity a suitable first introduction to astronomical photometry , and they have a good chance of being available in a high school setting . time spent getting to know these tools is well invested , as they can be used for numerous types of direct astronomical observation , as well . data analysis is simple , using variations of the inverse square law , modified by ratios of physical quantities . the calculations themselves are sufficiently simple as to be suitable for a microsoft excel or libreoffice calc spreadsheet . the street chosen was _ albert - fritz - strae _ in heidelberg , germany ; the beginning of the street can be found on google maps via the link [ http://www.google.com/maps/@49.382752,8.663361,19z ] . in this view , the street lights are on the upper pavement ; numbers in the text have been assigned from left to right , starting with the street s second light . measurements can be made using google s measuring tool in the classic map mode and may require the user to be logged in . remember to uncheck the `` 45@xmath32 '' option for a straight - down view . for the image , the camera was positioned right below the top left corner of the roof of first house in the street , on the lower side in the standard google view . the lanterns are visible as white circles on the upper pavement ( and cast shadows toward the top right ) . images and supplementary material ( including an excel worksheet and imagej macros documenting the photometric measurements ) can be found on [ http://www.haus-der-astronomie.de/materials/distances/street-lights ] . the photometric image for this example was taken with a canon eos 70d with 1855 mm canon kit lens at 55 mm , exposure 1/160 s at f/8.0 and iso 200 . lower iso values make for lower image noise ; the f stop should be chosen in a way to allow for the street lights at their different distances all to be in focus , and exposure should be such that no pixel is saturated . ( a quick way of checking is to plot a profile of pixels across the brightest lamp in the image ; if that profile has a flat top region , there is likely to be saturation . ) the raw ( .cr2 ) files from the camera were transformed into fits files using fitswork [ http://www.fitswork.de ] ( which in turn uses dcraw [ http://www.cybercom.net/~dcoffin/dcraw/ ] for the conversion ) , and brightness measurements were made with the polygon tool in imagej [ http://imagej.nih.gov/ij/ ] . all brightness measurements made in imagej were recorded as macros , for reference ; the macro files are available as part of the supplementary material available for download . imagej measurements can be saved as tab - separated files ( which , confusingly , will get a .xls extension ; when opened with microsoft excel or libreoffice calc , these files will need to be imported ; if you are in a country that uses decimal commas instead of decimal points , make sure that decimal places get imported correctly ) . the image showing extinction by tree branches was taken with the same canon eos 70d , using a tamron 18 - 200 mm zoom lens at 200 mm . standard candles and the inverse - square law are key concepts in astronomy . both concepts are necessary to understand modern determination of distances beyond our solar system ; such distance measurements , in turn , are a necessary prerequisite to deducing certain physical properties - most importantly intrinsic luminosity - from observations of astronomical objects . in the uk , the astronomy gcse qualifications prepared by edexcel @xcite , which complements the key stage 4 ( ks4 ) science curriculum ( gsce science and gsce additional science ) include the understanding of apparent and absolute magnitude , their relation to distance , the inverse - square law and its use with cepheid standard candles in topic 3.3 . `` physical properties of stars '' of unit 1 `` understanding the universe '' ( 5as01 ) . skills obtained in the activity described above , namely photometric measurements using a digital camera and image - processing software for analysis , can be re - used for some of the observation activities in unit 2 `` exploring the universe '' ( e.g. observation tasks b7 , b9 ) . the scottish `` curriculum for excellence '' lists standard candles as part of their higher physics curriculum area `` our dynamic universe '' : under the heading `` the expanding universe '' , the course unit support notes explicitly list `` consideration of parallax measurement and the data analysis of apparent brightness of standard candles in measuring distances to distant objects '' @xcite . in germany , astronomy is a mandatory part of the high school curriculum in some of the country s federal states . in @xcite , the basic understanding of astronomy ( section 7.1 ) requires understanding of light as a carrier of information , and of receivers such as photographic cameras or ccds ; mandatory content for understanding stars and stellar systems ( section 7.6 ) includes brightness and distances of the stars , methods of distance determination , and the concept of absolute brightness . astrophotography is a suggested project activity ( appendix 3 ) . in sachsen - anhalt ( saxony - anhalt ) , the obligatory 9th grade curriculum for high schools ( gymnasium ) includes a segment `` stars and star system '' ( thema 5 ) , part of which is an experimental demonstration of the relationship between distance and apparent brightness @xcite . the goals of the general 10th grade astronomy curriculum include an understanding of the relationship between luminosity , apparent brightness and distance @xcite . for grades 11/12 , the curriculum for astronomy courses includes ( kurs 2 ) the relationship between apparent magnitude , absolute magnitude , and brightness @xcite . in thringen ( thuringia ) , 10th grade astronomy includes an understanding of the difference between apparent and true brightness as well as of at least one method of determining astronomical distances . methods for determining astronomical distances , including photometric distances and different types of standard candle , are part of the 12th grade curriculum ( themenbereich 3.2 ) . more generally ( themenbereich 3.1 ) , students are expected to be competent in making their own astronomical observations , including the use of digital photography and software for image processing and analysis @xcite . in bavaria , astrophysics is one of two possible themes for teaching physics in grade 12 , the other being quantum physics @xcite . in the section on stars ( ph@xmath33 12.4 ) , students are required to learn about apparent and absolute magnitudes as well as the distance modulus ; in the section on large - scale cosmic structure ( ph@xmath33 12.5 ) , about standard candles , notably cepheids and supernovae . natbib coble , k. et al . 2013 : http://esoads.eso.org/abs/2013aedrv..12a0102c[``investigating student ideas about cosmology i : distances and structure '' ] in _ astronomy education review _ v. 12 , n. 1 . doi : http://dx.doi.org/10.3847/aer2012038[10.3847/aer2012038 ] de grijs , r. 2011 : _ an introduction to distance measurement in astronomy . _ wiley & sons feast , m. w. 2013 : http://esoads.eso.org/abs/2013pss5.book..829f[``galactic distance scales '' ] in t. d. oswalt , g. gilmore ( eds . ) : _ planets , stars , and stellar systems . volume 5 : galactic structure and stellar populations_. springer : dordrecht . feast , m. w. 2005 : `` the current status of primary distance indicators '' in a. p. fairall , p. a. woudt , _ nearby large - scale structures and the zone of avoidance_. asp conference series , vol . 329 , p. 241250 . http://arxiv.org/abs/astro-ph/0405440[arxiv:astro-ph/0405440 ] , doi : http://dx.doi.org/10.1007/978-94-007-5612-0_16[10.1007/978-94-007-5612-0_16 ] fernie , j. d. 1969 : http://adsabs.harvard.edu/abs/1969pasp...81..707f[``the period - luminosity relation : a historical review '' ] in _ publications of the astronomical society of the pacific _ vol . doi : http://dx.doi.org/10.1086/128847[10.1086/128847 ] gregory , matthew : _ specification : edexcel gsce in astronomy ( 2as01 ) for first certification 2014 , issue 2 . _ pearson education limited 2011 . mecklenburg - vorpommern , ministerium fr bildung , wissenschaft und kultur 2004 : http://www.bildung-mv.de/schueler/schule-und-unterricht/faecher-und-rahmenplaene/rahmenplaene-an-allgemeinbildenden-schulen/astronomie/[rahmenplan astronomie - jahrgangsstufen 7 - 10 ] ( last accessed on 2017 - 02 - 08 ) phillips , m. m. 1993 : http://adsabs.harvard.edu/doi/10.1086/186970[``the absolute magnitudes of type ia supernovae '' ] in _ astrophysical journal letters _ vol . doi : http://dx.doi.org/10.1086/186970[10.1086/186970 ] sachsen - anhalt , kultusministerium 2003 : http://www.bildung-lsa.de/index.php?historyback=1&kat_id=2171[rahmenrichtlinien gymnasium astronomie schuljahrgang 9 , wahlpflichtfach : schuljahrgnge 9 - 12 ] ( last accessed on 2017 - 02 - 08 ) sachsen - anhalt , kultusministerium 2012 : http://www.bildung-lsa.de/index.php?historyback=1&kat_id=2171[fachlehrplan sekundarschule : astronomie ] ( last accessed on 2017 - 02 - 08 ) schmidt , b. p. 2012 : http://esoads.eso.org/abs/2012rvmp...84.1151s[``nobel lecture : accelerating expansion of the universe through observations of distant supernovae '' ] in _ reviews of modern physics _ , vol . 84 , issue 3 , pp . doi : http://dx.doi.org/10.1103/revmodphys.84.1151[10.1103/revmodphys.84.1151 ] scottish qualifications authority 2014 : http://www.sqa.org.uk/sqa/47916.html[_unit support notes physics : our dynamic universe ( higher ) _ ] ( last accessed on 2017 - 02 - 08 ) staatsinstitut fr schulqualitt und bildungsforschung mnchen 2004 : http://www.isb-gym8-lehrplan.de/contentserv/3.1.neu/g8.de/index.php?storyid=27147[_lehrplan fr das gymnasium in bayern physik jahrgangsstufen 11/12 _ ] ( last accessed on 2017 - 02 - 08 ) stanger , jeff 2008 : `` demonstrations illustrating the difficulties astronomers face when observing astronomical objects '' in _ astronomy education review _ v. 7 n. 2 , pp . 132 - 140 . http://dx.doi.org/10.3847/aer2008033[doi : 10.3847/aer2008033 ] thringer ministerium fr bildung , wissenschaft und kultur 2012 : https://www.schulportal-thueringen.de/web/guest/media/detail?tspi=3122[_lehrplan fr den erwerb der allgemeinen hochschulreife : astronomie _ ] ( last accessed on 2017 - 02 - 08 ) webb , s. 1999 : _ measuring the universe : the cosmological distance ladder . _ springer : berlin , heidelberg , new york . weinberg , s. 2008 : _ cosmology_. oxford university press : oxford .	astronomers measure cosmic distances to objects beyond our own galaxy using standard candles : objects of known intrinsic brightness , whose apparent brightnesses in the sky are then taken as an indication of their distances from the observer . in this activity , we use street lights and a digital camera to explore the method of standard candles as well as some of its limitations and possible sources of error .
in classical geometry and physics , the geometry of a homogeneous space and its isometry group mutually determine each other . any deformation of either the geometry into a non - commutative version or of the isometry group into a hopf algebra should preserve as much of this interplay as possible . one way of achieving this , at least at the algebraic level , is to combine the space and its isometry into one algebraic object . in physics terms , such an algebra should combine position coordinates for space with generators of rotations and translations , i.e. , with angular momenta and momenta . momenta and positions are in duality and rotations act on both . this is manifest in three important subalgebras : the heisenberg algebra generated by positions and momenta , the isometry algebra generated by angular momenta and momenta and finally the algebra of angular momenta and positions , which one may interpret as an isometry algebra of momentum space . in these general terms , semiduality is a bijection between the isometry algebras of space and momentum space . a mathematically precise version of these ideas was formulated by majid in terms of hopf algebras @xcite or , infinitesimally , lie bialgebras @xcite , and is summarised in the textbook @xcite . majid s approach is in turn inspired by born s proposal of a reciprocity between momenta and positions @xcite . here we will use majid s framework , working at the level of lie bialgebras . thus , non - commutativity of a space is parametrised by lie brackets of position coordinates and its curvature by their co - commutator . in the dual lie bialgebra , commutators and co - commutators are swapped so that curvature of space is also captured in the commutators of momenta ; similarly curvature of momentum space is captured in the co - commutators of momenta or the commutators of positions . curved , homogenous geometries and the associated lie brackets of isometry algebras are a standard topic in classical geometry and physics . curved , homogeneous momentum spaces and the associated co - commutators of momenta are less familiar but play a central role in 3d ( quantum ) gravity ( see @xcite for a review ) , in so - called @xmath1-deformations of the poincar lie algebra @xcite and more recently in the discussion about relative locality @xcite . in this paper we exploit our familiarity with curvature and non - commutativity on one hand to enhance our understanding of co - commutativity on the other . we consider a family of lie algebras , thinking of them as rotation - position algebras or infinitesimal isometries of momentum space , and systematically compute their semidual bialgebras . the semiduals have non - trivial co - commutators which are co - boundary , i.e. , given by a classical @xmath0-matrix . we derive a formula for the @xmath0-matrix in terms of the map which characterises the split of the original lie algebra into rotation generators and positions . specialising to three dimensions , the semiduals of our family of lie bialgebras have the lie brackets of the 3d euclidean or poincar groups . lie bialgebra structures on these lie algebras are necessarily co - boundary and the possible @xmath0-matrices were classified a while ago by stachura in @xcite . denoting angular momenta by @xmath2 and momenta by @xmath3 , the list found by stachura includes @xmath0-matrices which only contain rotation or lorentz generators ( these are just the standard @xmath0-matrices for @xmath4 ) , trivial solutions which only contain the commuting momentum generators , and then a more interesting and complicated list of solutions of the mixed form @xmath5 . our approach via semiduality reproduces all these mixed solutions and relates them to the lie brackets of the original algebra . in particular , we are able to relate these @xmath0-matrices to the bianchi classification of three - dimensional lie algebras . the work reported here is closely related to our previous paper @xcite in which we studied the semidual hopf algebras of the universal enveloping algebras of the isometry lie algebras arising in 3d gravity . here we generalise and extend the results of @xcite at the infinitesimal , lie bialgebra level . however , in comparing the current paper with @xcite and also with previous work on semiduality in 3d quantum gravity @xcite the reader should be aware of a possible confusion between different interpretations of semiduality . here we interpret semiduality essentially as the exchange of position and momentum generators , as outlined above . in @xcite , by contrast , pairs of semidual hopf algebras are both thought of as ` rotation - momentum ' algebras , but semiduality exchanges the regimes of 3d quantum gravity where they play the role of quantum isometry groups . the paper is organised as follows . in sect . 2 we introduce the family of lie algebras whose direct sum decomposition we would like to study . they are necessarily even - dimensional and obtained from a given real lie algebra @xmath6 by a process of generalised complexification . we parametrise decompositions of lie algebras in the original family as double cross sums with @xmath6 as one of the factors in terms of a map @xmath7 and derive a quadratic lie - algebraic relation for @xmath8 which ensures that it does indeed define a double cross sum . the main result of this section is that such maps also characterise classical @xmath0-matrices of the semidual lie bi - algebra , and that the quadratic relation obtained as factorisation condition is equivalent to the modified classical yang - baxter equation for the semidual lie bi - algebra . in sect . 3 we specialise to the case where @xmath6 is the lie algebra of either the rotation group in euclidean 3-space or the lorentz group in three spacetime dimensions . the resulting family of complexified lie algebras are the local isometry groups of 3d gravity @xcite , but here we think of them as the lie algebras of rotations or lorentz boosts together with position coordinates , so that , in keeping with the general philosophy explained above , their semiduals , which have the lie algebra structure of the 3d euclidean or poincar lie algebra , can be interpreted as spacetime symmetries . using the isomorphism between @xmath9 and @xmath6 in this case , we reformulate the quadratic lie - algebraic relation for @xmath8 as a quadratic equation for the linear map @xmath8 , and , using rotational or lorentz symmetry , project out three equations which have to be satisfied independently . solving these equations in sect . 4 we recover a family of @xmath0-matrices for the euclidean and poincar lie algebras first found by stachura @xcite , and show that they are in one - to - one correspondence with the types in the bianchi classification of three - dimensional lie algebras bar the heisenberg algebra . 5 contains a summary of our results in tabular form and a discussion . consider an _ _ arbitrary @xmath10-dimensional real lie algebra @xmath6 with generators @xmath11 and brackets = f_ab^cj_c . here and in the following we use the einstein summation convention . picking a real number @xmath12 , one can associate to this lie algebra the following @xmath13-dimensional real lie algebra @xmath14 with generators @xmath15 , additional generators @xmath16 , @xmath17 , and brackets of the cartan form [ genisolie ] [ j_a , j_b]=f_ab^cj_c , = f_ab^c_c , = f_ab^cj_c . for @xmath18 this is the usual complexification @xmath19 , for @xmath20 it is the semidirect product @xmath21 and for @xmath22 it is isomorphic to @xmath23 . one can unify these three cases by considering the lie algebra @xmath14 as a generalised complexification with a formal parameter @xmath24 which satisfies @xmath25 and the identification @xmath26 , see @xcite and @xcite . we will not emphasise this viewpoint in the following , but will make use of the linear map @xmath24 defined via : _ _ , j_a_a , _aj_a a=0, ,n . lie algebras like arise in physics , particularly in three spacetime dimensions . in that context , the generators @xmath2 are rotation- or lorentz generators , the @xmath16 are usually interpreted as components of momentum , i.e. , as generators of spacetime translations , and the constant @xmath27 is related to the cosmological constant . in the current context we also think of the @xmath2 as generalisations of rotation or lorentz generators . however , when we apply semiduality we switch from the generators @xmath16 to generators of the dual vector space , which we then interpret as momentum space . from that point of view , the generators @xmath16 should be thought of as position coordinates , and @xmath12 as a real constant related to the curvature of momentum space @xcite we are interested in all decompositions of @xmath14 as double cross sum @xmath28 of @xmath6 and a second @xmath10-dimensional lie algebra @xmath29 . we can assume without loss of generality that the generators of @xmath30 are of the form [ genguessgen ] _ a=_a + ^b_aj_b , a=1, ,n , for a real @xmath31 matrix @xmath32 . as explained in @xcite , the most general form of the brackets in such a double cross sum is [ gennisolie ] [ j_a , j_b]=f_ab^cj_c , = f_ab^c_c + l_ab^c j_c , = g_ab^c_c . if the matrix elements @xmath32 can be found so that the generators @xmath33 close under lie brackets then we can express the structure constants @xmath34 and @xmath35 in terms of @xmath36 and @xmath37 as follows : @xmath38 to determine the condition on the generators to form a lie subalgebra we think of @xmath32 as the matrix relative to the basis @xmath11 of a linear map : , x = x^a j_a^b_ax^a j_b , then we define the map [ idef ] i : _ , xx + ( x ) , so that the sought - after generators in are [ genguessgenn ] _ a = i(j_a ) , a=1, ,n . using this notation , we show the following . the condition on @xmath39 for the generators to form a lie subalgebra of @xmath14 is [ mcond ] [ ( x),(y)]-([x,(y)]+[(x),y ] ) = - x , y . * proof * in terms of the map we need to find the condition under which the commutator of the two elements of the form @xmath40 and @xmath41 lies in the image of @xmath42 . one computes = + [ ( x),(y ) ] + ( [ x,(y)]+ [ ( x),y ] ) . for this to be of the form @xmath43 we require ( [ x,(y)]+ [ ( x),y])=+[(x),(y ) ] , as claimed . @xmath44 note that , when the map @xmath36 solving can be found , then the map @xmath42 is a bijection @xmath45 the process of semidualisation is defined for lie bialgebras which are double cross sums , see @xcite for a systematic treatment . we are going to apply it to the lie bialgebras @xmath46 , with the lie algebra structure already discussed and trivial co - commutator . in the semidual lie bialgebra , the lie algebra @xmath30 is replaced by its dual lie bialgebra @xmath47 . in our case , @xmath30 has trivial co - commutators so that the lie algebra structure of @xmath47 is abelian . following the notation of @xcite , we denote the semidual lie bialgebra of @xmath46 by @xmath48 in order to give its commutators and co - commutators , we introduce generators @xmath49 , @xmath50 of @xmath47 which are dual to the generators @xmath33 of @xmath30 , i.e. , they satisfy ^a(_b)=^a_b . then , in terms of the generators @xmath51 of the semidual lie bialgebra , the commutators are [ semid ] [ j_a , j_b]=f_ab^cj_c , = -f_ac^b^c , = 0 , or @xmath52=f_{ab}^{\;\;\;c}j_c , \quad [ \cp^a , j_b]=f_{bc}^{\;\;\;a}\cp^c , \quad [ \cp^a,\cp^b]=0.\ ] ] applying the general formulae in sect . 8.3 of @xcite , the co - commutators in our case come out as @xmath53 for each double cross sum decomposition of the lie algebra @xmath14 in the form @xmath46 and with generators of @xmath30 given in terms of the map @xmath39 by , the semidual lie bialgebra @xmath54 has a co - commutator which is co - boundary , with classical @xmath0-matrix [ rmatrix ] r = p^a f ( j_a ) = ^b_a^a j_b . * proof * consider the following ansatz for an @xmath0-matrix of @xmath54 [ genr ] r = r^b_a ^a j_b . for the associated co - commutators one finds @xmath55 so that we obtain the relation @xmath56 comparing with the expressions for @xmath35 and @xmath57 given in we conclude that we can reproduce them by choosing [ result1 ] r^a_b=^a _ b. finally , we determine a condition for to satisfy the modified classical yang - baxter equation [ mybe ] [ [ r , r]]+ = 0 , where , for now , @xmath12 is an arbitrary real constant and @xmath58 is the invariant element = f_ab^c ( ^a ^b j_c - ^a j_c ^b + j_c ^a ^b ) . inserting into = [ r_12,r_13]+ [ r_12,r_23 ] + [ r_13,r_23 ] yields three terms , ( r^b_a r^c_df_be^d - r^b_e r^c_df_ba^d+r^b_e r^d_af_bd^c ) ^e ^a j_c and similar terms proportional to @xmath59 , and @xmath60 . we deduce that the modified yang - baxter equation is equivalent to r^b_a r^c_df_be^d - r^b_e r^c_df_ba^d+r^b_e r^d_af_bd^c+f_ea^c=0 . however , this is simply the matrix form of the equation for the map @xmath36 . thus , substituting into we indeed obtain a solution of the modified classical yang - baxter equation . @xmath44 we now turn to solutions of the factorisation condition for isometry lie algebras arising in 3d gravity . in the following @xmath6 stands for either @xmath61 or @xmath62 , with generators @xmath2 , @xmath63 . this range of indices is unconventional in the euclidean setting but well - adapted to the more intricate lorentzian situation . we write @xmath64 for either the euclidean metric diag@xmath65 or the lorentzian metric diag@xmath66 , and use it lower or raise indices . the lie brackets of @xmath6 are then [ so123 ] [ j_a , j_b]=_abcj^c , where we adopt the convention @xmath67 . the invariant inner product [ innerprod ] j_a , j_b=_ab on @xmath6 plays an important role in our analysis . in the lorentzian case , it is sometimes convenient to work with normalised raising and lowering operators [ newgens ] n=(j_0+j_2 ) , n = ( j_0-j_2 ) , whose names are chosen to reflect the fact that they are null ( or lightlike ) with respect to : n , n = n , n = 0 , n , n = 1 . their commutators are [ n , n ] = j_1 , = n , = -n . the lie algebra @xmath14 is that of the poincar , de sitter or anti - de sitter group or their euclidean analogues in three dimensions , with brackets [ isolie ] [ j_a , j_b]=_abcj^c , = _ abc^c , = _ abc j^c . as in the general case , we are looking for basis change [ guessgen ] _ a=_a + ^b_aj_b , so that @xmath68 is a basis and @xmath69 closes under lie brackets , i.e. , [ eup ] [ j_a , j_b]=_ab^cj_c , = _ ab^c_c + l_ab^c j_c , = g_ab^c_c . in this case , the semidual lie algebra has the brackets [ eucllorentz ] [ j_a , j_b]=_ab^cj_c , = _ ab^c^c , = 0 , which are the brackets of the euclidean lie algebra in the case of euclidean signature and those of the poincar lie algebra in the case of lorentzian signature . every decomposition of according to will therefore lead to a co - boundary lie bialgebra structure on those lie algebras . in order to determine all possible solutions of the condition , we will be making use of the invariant inner product on @xmath6 . we write @xmath70 for the transpose of a map @xmath39 relative to @xmath71 , i.e. , for the map which satisfies ^t ( x),y= x , ( y)x , y . we also need the fundamental identity [ epsid ] _ abc^efg = _ a^e ( _ b^f _ c^g -_b^g _ c^f ) - _ a^f ( _ b^e _ c^g -_b^g _ c^e ) + _ a^g ( _ b^e _ c^f -_b^f _ c^e ) , which holds in both the euclidean and lorentzian context . it implies _ abc^afg= _ b^f _ c^g -_b^g _ c^f , which , in turn , is equivalent to [ special ] [ x,[y , z]]=x , zy -x , yz , x , y , z . this can be used to prove the useful result [ use1 ] , vv , z+ , vv , x + , vv , y = v , v , z for any @xmath72 , see @xcite for details and a related identity . the lie bracket or , equivalently , the epsilon tensor provide an identification of @xmath73 with @xmath6 . it follows that for any linear map @xmath39 , the assignment ( x , y)defines a linear map @xmath74 . our factorisation condition allows for a convenient ` dual ' formulation in terms of this map , which turns out to be the adjugate of @xmath36 : for every linear map @xmath75 , there is a uniquely determined linear map ^adj : , which satisfies [ deftau ] ^adj(z),[x , y]= z,[(x),(y)]x , y , z . it is given by [ tmaster ] ^adj= ^2-tr ( ) + 1 2 ( tr())^2-tr(^2))id , which is the adjugate of @xmath36 . * proof * inserting the basis elements @xmath76 for @xmath77 in , one deduces the matrix relation ( ^adj)^f_e _ fab=_ecd ^c_a^d_b . multiplying with @xmath78 , summing over @xmath79 and repeatedly applying one arrives at the formula . to see why gives the adjugate of @xmath36 ( usually defined as the transpose of the matrix of co - factors ) note that for any @xmath31 matrix @xmath80 , the characteristic polynomial @xmath81 has the constant term @xmath82 so that q_a(t)= is a polynomial of degree @xmath83 which satisfies q_a(a ) a= aq_a(a)= ( -1)^n-1 ( a ) i d , by the cayley - hamilton theorem . it is proved in @xcite that , in fact , a^adj = q_a(a ) . the expression is easily seen to be @xmath84 for @xmath85 so that the solution of is indeed the adjugate of @xmath36 as claimed . @xmath44 note that for @xmath86 or @xmath87 , the invariance of the epsilon tensor implies @xmath88 . inserting this into is simply the cayley - hamilton theorem for the @xmath89 matrix @xmath90 . in the case @xmath91 or @xmath92 , the factorisation condition for the linear map @xmath39 is equivalent to the quadratic relation [ mconddd ] ( -tr id)(+^t)+1 2 ( ( tr)^2-tr(^2))id =- id * proof * the factorisation condition can equivalently be written as @xmath93- \langle \mm^t ( z),[x,\mm(y)]+[\mm(x),y]\rangle \rangle \qquad \qquad \nonumber \\ = -\lambda\langle z , [ x , y]\rangle \quad \forall \ ; x , y , z \in{\mathfrak{g}}.\end{aligned}\ ] ] now observe that + [ ( x),y]=[(id+)(x),(id+)(y)]-[(x),(y)]-[x , y ] , and apply both to @xmath36 and @xmath94 to deduce . @xmath44 in order to analyse the ( equivalent ) conditions and further , we split @xmath36 into symmetric and antisymmetric parts with respect to @xmath71 . the antisymmetric part can be written as the adjoint action of a general element @xmath95 by virtue of the identity , z+ y,[x , z]=0 x , y , z . thus we write the map @xmath36 as [ split ] = s + ad_v . for @xmath96 and @xmath97 satisfying s(x),y= x , s(y ) x , y . inserting the split for @xmath36 and simplifying using and standard identities we deduce @xmath98\rangle - \langle x,[s(y),s(z)]\rangle - \langle y,[s(z),s(x)]\rangle \qquad \qquad \qquad \nonumber \\ + \langle 2 [ v , s(z)]+(\lambda + \langle v , v\rangle ) z , [ x , y]\rangle = 0\quad \forall x , y , z \in { \mathfrak{g}},\end{aligned}\ ] ] which is equivalent to @xmath99-s([x , s(y)]+ [ s(x),y ] ) \qquad \qquad \qquad \qquad \qquad \qquad \nonumber \\ + 2(\langle v , x \rangle s(y ) - \langle v , y\rangle s(x ) ) = -(\lambda + \langle v , v\rangle)[x , y ] \quad \forall x , y \in { \mathfrak{g}}. \end{aligned}\ ] ] this turns out to be a very useful formulation of the factorisation condition for the lie algebras . we can derive an equivalent condition to by inserting the split into and using @xmath100 . a short calculation gives [ masterr ] 2s^2 - 2tr(s ) s+1 2 ( ( tr(s))^2-tr(s^2))id+ 2 ad_v s= - ( + v , v)id . this can also be derived directly from using the methods of sect . [ tricks ] . the equation is invariant under @xmath101 or @xmath102 conjugation , and this can be used to split it into irreducible components under this action , namely into symmetric traceless , antisymmetric and scalar matrices . noting that the commutator @xmath103 $ ] is symmetric but that the anticommutator @xmath104 is antisymmetric we write 2 ad_v s = [ ad_v , s]+\{ad_v , s } and deduce three equations . for the scalar part we have [ scalar ] 1 6 ( ( trs)^2 - tr(s^2 ) ) = + v , v , for the vector part we find [ vect ] \{ad_v , s } = 0 and for the symmetric , traceless part we deduce [ symtr ] [ ad_v , s]+ 2(s^2 - 1 3 tr(s^2)id ) -2(trs s -1 3 ( trs)^2id)=0 . the equations , and are derived in a different form and by a different method in @xcite , where they are used for determining the most general @xmath0-matrix for the poincar and euclidean lie algebras . adding and we deduce the relation [ handy ] ad_v s+ ( s^2 - 1 3 tr(s^2)id ) -(tr(s ) s -1 3 ( trs)^2id)=0 . this has a various useful consequences , including the following lemma which is also implicit in some of the manipulations in the appendix of @xcite . [ neat ] if the map @xmath39 solves the factorisation condition and @xmath105 and @xmath106 are its symmetric and antisymmetric part as in , then the following hold : 1 . the antisymmetric part @xmath106 is the zero map on any subspace of @xmath107 where the restriction @xmath108 is invertible . in particular , if @xmath105 is invertible then @xmath109 vanishes . if dim ker @xmath105=1 then : ker @xmath105 is not a null space @xmath110 @xmath111 . in the following , the contrapositive of result ( 2 ) will be particularly useful : if ker @xmath105 is one - dimensional and @xmath112 then ker @xmath105 is necessarily a null space . * proof * ( 1 ) on any subspace @xmath113 where @xmath114 is invertible we can multiply by the ( symmetric ) map @xmath115 to obtain an expression for the antisymmetric map @xmath106 in terms of a symmetric map . thus @xmath116 . \(2 ) it follows from that ker @xmath105 is invariant under the action of @xmath106 . if ker @xmath105 is one - dimensional then any basis vector @xmath117 of it is necessarily an eigenvector of @xmath106 , i.e. , @xmath118=\mu x$ ] . since @xmath119 , x\rangle = 0 $ ] , we deduce that @xmath120 if @xmath117 is not null . however , in that case , @xmath105 restricted to the orthogonal complement of @xmath117 is invertible , so we know from part ( 1 ) that @xmath106 vanishes on the orthogonal complement of @xmath117 . since we already have @xmath121 we deduce that @xmath122 . @xmath44 before we systematically study solution of the equation in the next sections , we note some special cases which can be found by inspection . setting the symmetric part @xmath105 to zero , the condition reduces to a condition on the lie algebra element @xmath109 . expanding [ vexp ] v =- v^aj_a , where the sign is chosen to match conventions in @xcite , we have [ vcond ] v , v= v^a v_a = - . the resulting bialgebra structure on the semidual lie algebra has the @xmath0-matrix [ bicr ] r_=v^c ^b _ ac ^a j_b , which is the familiar 3d @xmath1-poincar structure with deformation parameter @xmath123 , which may be timelike , spacelike or timelike , depending on @xmath12 @xcite . the resulting lie algebra structure of @xmath30 is a semidirect sum @xmath124 , with @xmath125 acting on @xmath126 by scaling . we will recover this solution as part of a more general family in the next section . similarly , setting the antisymmetric part @xmath106 of the map @xmath36 to zero , we deduce from equations and that [ vzero ] s^2-tr(s ) s = - 2 i d v=0 at the same time , we know from the cayley hamilton theorem that s^3-tr(s ) s^2 + 1 2 ( ( tr(s))^2-tr(s^2))s-(s)id=0 . thus , from with @xmath122 ( s^2-tr(s ) s + 3id)s-(s)id=0 . combining with we deduce s = s i d v=0 . it follows that for @xmath127 any purely symmetric solutions is diagonal and given by [ diagsol ] = i d . the resulting classical @xmath0-matrix [ doubler ] r_= ^a j_a is that of the classical double of @xmath61 or @xmath62 . the associated lie brackets on @xmath30 are = 2_abc^c , i.e. , @xmath30 is @xmath61 or @xmath62 in this case . when @xmath128 we have another symmetric solution that can be found by inspection ( but which we will recover systematically later ) . for any vector @xmath129 , the liner map with matrix [ trivsol ] ^b_a = m^bm_a . trivially satisfies ( for @xmath20 ) , so that r_m = m_a m^b ^a j_b is a classical @xmath0-matrix for the euclidean or the poincar lie algebras . the lie algebra structure of @xmath30 turns out to be a semidirect sum @xmath124 with @xmath125 acting on @xmath126 by rotations , lorentz boosts or the nilpotent matrix [ jord ] m = 0 & 1 + 0 & 0 , depending on whether @xmath129 is timelike , spacelike or lightlike . we will study this solution , too , as part of a larger family in the next sections and give more details there . in the simplest case , the symmetric part @xmath105 has a diagonalising orthonormal basis @xmath130 , @xmath63 , of @xmath6 with real eigenvalues @xmath131 , @xmath63 . the diagonalisation is always possible in the euclidean case but can not always be achieved in the lorentzian case . we will consider the other , non - standard cases below . since the diagonalising basis is related to the original basis @xmath2 , @xmath63 by an orthogonal ( i.e. , @xmath101 or @xmath102 ) transformation , we know that the commutators are still = _ abc k^c . the map @xmath105 in then takes the from [ ansatz ] s=_a=0 ^ 2 _ a k_ak_a , . in order to find all solutions of this form , we consider two cases . + ( i ) dim ker@xmath132 . in that case , it follows from lemma [ neat ] that @xmath122 : if @xmath105 is invertible this is a consequence of part ( 1 ) while for a one - dimensional kernel of @xmath105 the vanishing of @xmath105 follows from part ( 2 ) since ker @xmath105 can not be a null space in the diagonalisable case under consideration . with @xmath122 we insert the diagonal form of @xmath105 into to find @xmath133 taking pairwise differences we deduce _ 0=_1=_2=. this is the solution already found by inspection . since the eigenvalues @xmath131 are assumed to be real , this is only a valid and non - trivial solution if @xmath127 . \(ii ) dim ker @xmath134 . in that case , @xmath135 , so that the scalar equation gives the condition [ vcondd ] v , v= - for @xmath136 and the matrix condition reduces to = 0 . if @xmath137 then this imposes no further restriction of @xmath109 and we recover the purely antisymmetric solution @xmath138 found by inspection in the previous section . however , if dim ker @xmath105=2 so that @xmath139 for some fixed @xmath140 and @xmath141 , we see that we can now choose @xmath109 to be a multiple of @xmath130 , with the multiple chosen so that the normalisation condition holds . when @xmath142 , the resulting solution can be written [ firstnew ] = vv , + ad_v , with @xmath143 arbitrary and @xmath109 satisfying . when @xmath20 , the requirement that @xmath139 ( no sum ) and @xmath106 commute enforces @xmath122 . in that case we recover the solution , at least for space- or timelike vectors @xmath129 . the case of light - like @xmath129 represents a non - diagonalisable map @xmath105 and will appear in sect . [ nonstandard ] . in the remainder of this subsection we study the solution in more detail . with the convention , the generators of the subalgebra @xmath30 in this case are [ newgen ] _ a=_a + _ abcv^bj^c + ( v^bj_b ) v_a , where @xmath144 is an arbitrary ( real ) parameter . using 3-vector notation @xmath123 etc . we can also write [ newgenn ] = + + ( ) . then , with @xmath145 etc . , the brackets are @xmath146&\!\!\!=\!\!\!\!&-\bv\times(\bp\times \bq)\cd(\bpp-\beta \bv\times \bpp ) \nonumber \\ & \!\!\!=\!\!\!\!&(\bv\cd\bp)(\bq\cd\bpp)-(\bv\cd\bq)(\bp\cd\bpp ) -\beta ( \bp\times\bq)\cd\bv \ ; ( \bv\cd\bpp ) - \lambda\beta(\bp\times\bq)\cd\bpp\end{aligned}\ ] ] as well as [ subliepj ] [ , ] = + ( ) + ( ) , which can be written in terms of the generators @xmath33 as @xmath147 & = \bp\times \bq\cd\bpp+ ( \bv\times \bp -\beta \bv \times ( \bv \times \bp))\cd \bq\times \bj . \nonumber \\ & = \bp\times \bq\cd\bpp + ( \bv\cd \bq)(\bp\cd\bj ) -(\bp\cd\bq)(\bv\cd \bj ) -\beta(\bp\cd\bv)(\bv\times\bq\cd \bj)- \lambda\beta ( \bp\times\bq\cd \bj ) .\end{aligned}\ ] ] to identify the resulting lie algebra structure of @xmath30 , we consider the various cases . we pick @xmath149 , so that gives _ 0 = _0-j_0,_1=_1-j_2,_2=_2+j_1 . then yields the following brackets for @xmath30 : @xmath150&=0,\nonumber \\ [ \pp_0,\pp_1]&=\sqrt{-\lambda}\pp_1-\beta \lambda \pp_2,\nonumber \\ [ \pp_0,\pp_2]&=\sqrt{-\lambda}\pp_2+\beta \lambda \pp_1.\end{aligned}\ ] ] thus @xmath151 and @xmath152 span a commutative subalgebra , and @xmath153 acts on this by infinitesimal scaling and rotation . in terms of the bianchi classification of three - dimensional real lie algebras , this is type vii . the commutators involving @xmath2 and @xmath154 are @xmath155&= \pp_2 , \quad [ j_0,\pp_2 ] = -\pp_1,\nonumber \\ [ j_1,\pp_0]&= \sqrt{-\lambda } j_1+\beta \lambda j_2 -\pp_2 , \quad [ j_2,\pp_0 ] = \sqrt{-\lambda } j_2-\beta \lambda j_1 + \pp_1 , \nonumber \\ [ j_1,\pp_2]&= \pp_0- \beta \lambda j_0 , \quad [ j_2,\pp_1 ] = - \pp_0+\beta \lambda j_0 , \nonumber \\ [ j_0,\pp_0]&= 0 , \quad [ j_1,\pp_1 ] = [ j_2,\pp_2]= \sqrt{-\lambda}j_0,\end{aligned}\ ] ] and again allow for a geometric interpretation . the action of @xmath2 on the vector @xmath156 is infinitesimal rotation around the @xmath140-th axis . the action of @xmath153 on the @xmath157-plane is infinitesimal scaling and rotation , as on the @xmath158-plane above . the action of both @xmath151 and @xmath152 on the @xmath157-plane is to map it onto @xmath159 , which is consistent with the fact that @xmath151 and @xmath152 commute . if @xmath161 is the timelike vector @xmath162 , then the generators ( [ newgen ] ) of @xmath30 are _ 0 = _0-j_0,_1=_1+j_2,_2=_2-j_1 , so that from ( [ subliep ] ) we obtain & = & 0 , + & = & _ 1+_2 , + & = & _ 2-_1 . again we have a two - dimensional abelian algebra spanned by @xmath163 , with @xmath153 acting by infinitesimal rotation and scaling . the bianchi type is again vii . computing the mixed commutators is straight forward and leads to a geometric interpretation analogous to the euclidean case , but we omit the details here . with the spacelike vector @xmath165 , we have from ( [ newgen ] ) that _ 0 = _0- j_2,_1=_1-j_1,_2=_2-j_0 , and therefore ( [ subliep ] ) gives [ spacelikepp ] & = & 0 , + & = & -_0-_2 , + & = & -_2-_0 . it is illuminating to write the commutators in this case in terms of the generators @xmath166 defined in and the corresponding generators _n=(_0+_2),_=(_0-_2 ) . then _ n=_n - n,_=_+ , _ 1=_1-j_1 , and therefore ( [ spacelikepp ] ) gives [ spacelikeppn ] & = & 0 , + & = & -(+)_n , + & = & ( - ) _ , this is again a semidirect sum @xmath124 and bianchi type vi : @xmath167 and @xmath168 span a two - dimensional abelian algebra , and @xmath151 acts on them with a matrix that has ( generically distinct ) real eigenvalues . in the lorentzian case , maps @xmath97 which are symmetric with respect to the non - degenerate symmetric form @xmath169 can not always be diagonalised . we now consider the different normal forms ( which are discussed , for example , in the textbook @xcite ) . in each case , the solution of the factorisation condition or , equivalently , eq . uses lemma [ neat ] and standard arguments from linear algebra ; some proceed along similar lines to the discussion in the appendix of @xcite . in this case the symmetric map @xmath105 has a lightlike and a spacelike eigenvector and can be brought into the form s = a ( n n , + n n , ) + b nn , - c j_1j_1 , , b0 , so that s(n)= a n , s(n)= bn + a n , s(j_1)= cj_1 . it is easy to check that , if @xmath170 and @xmath171 ( so that @xmath105 is invertible ) we have @xmath122 by lemma [ neat ] and hence the diagonal solution . if @xmath170 then @xmath105 has a one - dimensional kernel which is not null , so by lemma [ neat ] we again deduce @xmath122 and obtain a solution of the form . if @xmath172 but @xmath171 we have , from , that @xmath173 . the requirement that @xmath106 leave the kernel of @xmath105 invariant implies that @xmath109 is spacelike or lightlike , i.e. , @xmath174 . when @xmath127 , we can use to deduce that @xmath175 . renaming @xmath176 as @xmath144 we have the new solution [ smalljordan ] = nn , -j_1j_1 , + ad_j_1 , > 0 , , so that _ n= _n+n,_=_+n- , _ 1=_1+j_1 . the commutators of @xmath30 are [ spacelikeppnb ] & = & 0 , + & = & 2 _ n , + & = & 0 . this is bianchi type iii , which is isomorphic to the direct sum of the one dimensional lie algebra @xmath125 ( generated here by @xmath177 ) and the non - abelian two - dimensional lie algebra @xmath178 ( generated here by @xmath179 ) . when @xmath20 , we again use to obtain the solution [ lightjordan ] = nn , + ad_n , for an arbitrary @xmath143 . this may be viewed as a limit as @xmath180 of and is also a lightlike version of . the generators of @xmath30 are _ n=_n,_=_-j_1+n , _ 1=_1-n , and the commutators of @xmath30 are & = & 0 + & = & -_n + & = & -_1-_n . which are again those of a semidirect sum @xmath124 when @xmath181 , this is bianchi type v , with @xmath182 acting on the span of @xmath183 via the identity . when @xmath184 , it is bianchi type iv , with @xmath185 acting on the span of @xmath183 via @xmath186 , where @xmath187 is the nilpotent linear map with matrix . the symmetric map @xmath105 only has a single , lightlike eigenvector and can be brought into the form s= a ( n n , + n n,- j_1j_1 , ) + b ( n j_1 , + j_1 n , ) , b0 so that s(n)= a n , s(n)= a n + b j_1 , s(j_1)= - bn + a j_1 . again we require @xmath172 for a non - vanishing @xmath109 . then ker @xmath105 is one - dimensional and spanned by the lightlike generator @xmath188 ; this space is invariant under @xmath106 provided @xmath109 is a linear combination of @xmath188 and @xmath189 . however , since @xmath190 we can use to deduce that @xmath191 , which in turn implies that @xmath192 . thus we are in the case @xmath193 and , after re - naming @xmath194 , we have the solution =( nj_1,+ j_1n , -ad_n ) , which can be written more compactly as [ largejordan ] = j_1n , . the resulting basis of @xmath30 is _ n = p_n , _ = _ + j_1 , _ 1=_1 , and we obtain the commutators = 0 , = 0 , = _ this is again a direct sum of @xmath125 ( spanned by @xmath151 ) and the two - dimensional non - abelian lie algebra , i.e. , bianchi type iii the remaining normal form of a symmetric map @xmath97 is s = a ( n n , + n n , ) + b ( n n,- nn , ) - c j_1j_1 , which has the form of a ` rotation ' in the span of @xmath195 : s(n)= a n + b n , s(n)= -b n + a n s(j_1)= c j_1 . in this case we do not obtain a new solution : the cases dim ker @xmath132 can be dealt with as above . however , dim ker @xmath105=2 requires @xmath196 , in which case we recover a solution in the family we can unify a large family of solutions of the factorisation condition in the form [ prettygeneral ] f= vv , + ad_v , , \{0,1 } , v , v= - . this reduces to the purely symmetric solutions when @xmath197 and to the family , including the lightlike case , when @xmath198 . this factorisation map gives rise to the @xmath0-matrices [ kappagen ] _ = ( v^a v_b + v^c^b_ac)p^aj_b , which generalise the familiar @xmath1-poincar @xmath0-matrix ; we therefore call them generalised @xmath1-poincar solutions . -actions in the semidirect sums @xmath124 and their degenerate limit @xmath199 ; explanation in the main text , title="fig : " ] -actions in the semidirect sums @xmath124 and their degenerate limit @xmath199 ; explanation in the main text , title="fig : " ] -actions in the semidirect sums @xmath124 and their degenerate limit @xmath199 ; explanation in the main text , title="fig : " ] + -actions in the semidirect sums @xmath124 and their degenerate limit @xmath199 ; explanation in the main text , title="fig : " ] -actions in the semidirect sums @xmath124 and their degenerate limit @xmath199 ; explanation in the main text , title="fig : " ] the lie algebra structure of @xmath30 arising from the family is of the semidirect sum form @xmath124 for all values of the parameters . the @xmath125-action on @xmath126 has two components , corresponding to the parameters @xmath200 and @xmath144 being non - zero . the @xmath200-component ( for @xmath201 and @xmath202 ) is simple overall scaling , shown in the bottom left of fig . [ flowpics ] and associated with the standard @xmath1-poincar algebra . the @xmath144-component ( for @xmath203 and @xmath197 ) is more interesting and depends on the value of @xmath12 . its matrix is proportional to one of the @xmath204 matrices @xmath205 representing the relation @xmath206 : ( ) = in addition to the family we found the solution @xmath207 i d , which gives rise to the @xmath0-matrix of the classical double and @xmath208 or @xmath209 . finally , we have the exceptional , jordan - type solutions and . with the notation @xmath210 , the @xmath0-matrix associated to the factorisation map is [ smallr ] r_= p_nn + ( p_1j_1 + ^b_a1 p^aj_b ) , while the @xmath0-matrix associated to the solution is simply [ larger ] r _ the associated lie algebra structure of @xmath30 is @xmath211 , which may be viewed as a degenerate case of the semidirect sums @xmath124 with a diagonalisable @xmath125-action but one eigenvalue vanishing . the eigenvector for the non - vanishing eigenvalue is always the lightlike vector @xmath188 , but the eigenvector for the zero eigenvalue varies between the two cases and as a function of the parameter @xmath144 . the flow obtained by exponentiating such matrices is shown for a generic case in the bottom right of fig . [ flowpics ] . table [ summary ] summarises our results and discussion of the factorisation of the lie algebras and the associated @xmath0-matrices of the semiduals . as reviewed in the introduction , semidualisation gives all @xmath0-matrices of the form @xmath212 in this case . it is interesting that for each type of @xmath0-matrices there is an associated bianchi type for the lie algebra @xmath30 . all bianchi types except type ii ( the heisenberg algebra ) arise in this way . viewing the generators @xmath33 of the original algebra as spacetime coordinates and the dual generators @xmath3 in as momenta , the summary in table [ summary ] may thus also be viewed as a list of non - commutative spaces and associated non - co - commutative momentum spaces , parametrised by the relevant @xmath0-matrices . we thus obtain the unified picture of non - commutative geometries and isometry algebras promised in the introduction . there are several questions and topics for further research which follow from the results reported here . mathematically , one would like to understand more generally when semidualising families of lie algebras with a given double cross sum decomposition provides an effective way of finding @xmath0-matrices for the semidual lie bialgebra . it would also be interesting to apply the same technique to families of lie bialgebras with non - trivial co - commutators , so that the semidual has a more complicated lie algebra structure than the semidirect sums found here . in the physics literature , the standard @xmath1-poincar ( with a timelike deformation parameter ) and the classical double lie bialgebras in table [ summary ] have been much studied , the latter ( but not the former - see @xcite ) being related to 3d quantum gravity . it would be interesting to see if the generalised @xmath1-poincar @xmath0-matrices with @xmath203 and the jordan - type @xmath0-matrices and also have applications in real or toy models of mathematical physics . * acknowledgments * some of the work reported in this paper was carried out while prince k osei was supported by an ictp phd fellowship . we thank the perimeter institute for hospitality and support during the final stage of the project . s. majid , physics for algebraists : noncommutative and noncocommutative hopf algebras by a bicrossproduct construction , j. algebra , 130 ( 1990 ) 1764 . s. majid , matched pairs of lie groups associated to solutions of the yang - baxter equations , pac . j. math . 141 ( 1990 ) 311332 . j. lukierski , a. nowicki , h. ruegg and v. tolstoi , @xmath213-deformation of poincar algebra , phys . b264 ( 1991 ) 331338 . j. lukierski , a. nowicki and h. ruegg , new quantum poincar algebra and @xmath1-deformed field theory , phys . b293 ( 1992 ) 344352 . s. majid and h. ruegg , bicrossproduct structure of the @xmath1-poincar group and non - commutative geometry , phys . lett . b334 ( 1994 ) 348354 .	we study the interplay between double cross sum decompositions of a given lie algebra and classical @xmath0-matrices for its semidual . for a class of lie algebras which can be obtained by a process of generalised complexification we derive an expression for classical @xmath0-matrices of the semidual lie bialgebra in terms of the data which determines the decomposition of the original lie algebra . applied to the local isometry lie algebras arising in three - dimensional gravity , decomposition and semidualisation yields the main class of non - trivial @xmath0-matrices for the euclidean and poincar group in three dimensions . in addition , the construction links the @xmath0-matrices with the bianchi classification of three dimensional real lie algebras . empg-13 - 12 * classical @xmath0-matrices via semidualisation * 20 pt * prince k. osei * + department of mathematics + university of ghana , po box lg 25 , legon , ghana + [email protected] * bernd j. schroers * + department of mathematics and maxwell institute for mathematical sciences + heriot - watt university , edinburgh eh14 4as , united kingdom + [email protected] 24 august 2013 , udpated 8 october 2013 16 pt
there are many ways to shuffle a deck of cards . one of the most common is to split the deck into equal halves and then riffle or dovetail shuffle the cards back together , wherein the cards from the two halves interlace . a perfect shuffle of this type is one where the cards alternate perfectly between the two halves . there are two types of perfect shuffles , depending on which card ends up on top . these are called `` in '' and `` out '' shuffles and have been frequently described in the literature ( see , e.g. @xcite or @xcite ) . another way to think about this shuffling is we split the deck in two equal stacks and go through the stack from top to bottom using a rule about how to put the two cards into the newly shuffled stack . in and out shuffles correspond to the two rules where either we always put the card from the second half of the stack on top or we always put the card from the first half of the stack on top . in this paper we start the examination of a new kind of shuffling where our rule is to set an order on the labeling of the cards ( we allow for labels to be used multiple times in the deck ) and we again go through from top to bottom but now we let the order of the labeling on the cards determine which one goes on top . that is , starting with @xmath5 cards with labels from the ordered list @xmath6 , divide the stack of cards into @xmath2 stacks of @xmath7 , then take the top card off of each stack , sort the @xmath2 cards ( according to the order of the labels , where if the labels agree then the order they are sorted is unimportant ) and add them to the new stack . an example of this is shown in figure [ fig : sorting ] where the labels are @xmath8 and we let the heaviest weights `` sink down '' ( i.e. , go to the lower card height where we start counting from the top card down ) . to the deck @xmath9 . ] equivalently , this is the same as starting out with a list of @xmath10 labeled objects , @xmath11 and putting this into a @xmath12 matrix where we proceed by filling up the rows left to right and top to bottom , @xmath13 now take this matrix and sort the elements in each _ column _ according to the ordering of the labels , @xmath14 and finally concatenate the columns to form the new list . @xmath15 as an example if we have @xmath16 and start with @xmath17 cards labeled @xmath18 ( where we again let the 3 labels be ordered @xmath19 ) then we have @xmath20 this process can be repeated and so we have @xmath21 and now we see we have a stack which will repeat itself every two shuffling steps , or in other words a _ periodic _ stack . a stack which returns to itself after one shuffling step will similarly be called a _ fixed _ stack . in general , we can relate the shuffling to a directed graph where each of the possible @xmath22 stacks are the vertices and we put a directed edge between two vertices if shuffling one stack gives the other . since the outcome of our shuffling is uniquely determined by the order of the cards , each edge will only have one edge going out ( though it is possible many edges can go in ) . this immediately gives us the following . given any stack of cards after applying the shuffling procedure finitely many times we will settle into a stack which is periodic under the shuffling . there now arise several natural questions . for example , how do we find fixed / periodic stacks ? what periods are possible ? how many periodic / fixed stacks are there ? how long does it take for a stack to settle into a periodic stack ? in this paper we will answer some of these questions . in section [ sec : cycles ] we will introduce a weight function on the subscripts and show how to use this to represent the shuffling by a poset structure . in section [ sec : gcd ] we will show in the case when @xmath3 ( where @xmath0 ) the possible periods are all divisors of @xmath4 . in section [ sec : posets ] we will show how to adopt the poset structures to find posets that generate all fixed and periodic stacks . finally , in section [ sec : conclusion ] we will give some concluding remarks . we will throughout assume the number of cards is @xmath23 , where @xmath1 is the highest power of @xmath2 that divides @xmath24 , and @xmath25 is the size of the stacks we split @xmath24 into when shuffling . for simplicity we will focus on subscripts , i.e. , @xmath26 means @xmath27 . the key to understanding this shuffling is looking at a column which will ( for some @xmath28 ) consist of the terms @xmath29 and after sorting will be sent to @xmath30 for example , returning to the case @xmath17 and @xmath16 then we get the following . this shows , for example , @xmath31 in some order ( depending on the labels of the cards ) . to help us understand what is happening it is useful to weight the subscripts . a shuffling weight function on the subscripts is a map @xmath32 which satisfies the following two conditions for @xmath33 : * @xmath34 . * @xmath35 . the first condition says the weight of the entries in the column and the weight of the entries in a block it maps to are equal . the second condition says the weights are distinct and increasing in a block ( and combined with the first condition says the weights of the column are also distinct ) . as an example for @xmath17 and @xmath16 one possible weight function is given below . @xmath36 if we now take this weight function and replace all the entries in the above diagram with the corresponding weights then our diagram now becomes the following ( for which it is easy to check the two conditions are satisfied ) . we can use this weight function to map the subscripts to the subscripts in a bijective manner so the weight is preserved . for instance in the third column we have @xmath37 . since @xmath38 , @xmath39 and @xmath40 then we would have @xmath41 , @xmath42 and @xmath43 . repeating this for each column / block combination we can now break up the elements of @xmath44 into cycles , and we can place these cycles into a poset structure where the height of a cycle is the weight of a subscript in the cycle ( by construction they are all equal ) . so for our particular weight function we would end up with three levels with the following cycles . this diagram essentially represents half of the shuffling ( i.e. , how we move from columns to blocks in the concatenation ) . we also need to represent the other half of the shuffling , which is sorting among the columns . to do this we add edges between levels in the diagram connecting any two subscripts which appear in the same column . doing this gives us the following diagram ( which we will refer to as the _ shuffling poset _ ) . the important part about the shuffling poset is that the higher ordered cards ( or by analogy heaviest cards ) will always try to `` sink down '' , i.e. , the highest label in a column will want to map to the smallest possible @xmath45 value which corresponds to the lowest level it can reach , the second highest label will similarly go to the second lowest level it can reach and so on . in particular , we have the following . shuffling can be carried out completely in the shuffling poset by placing the corresponding cards in their position and carrying out the following two steps : * using only the edges _ between _ levels in the poset swap cards so no edge connects a higher ordered card at height ( or weight ) @xmath46 in the poset with a lower ordered card at height ( or weight ) @xmath47 where @xmath48 . * using only the edges _ in _ the levels advance the card to the next entry . we now see that once we have a shuffling weight function we can generate the shuffling poset which completely describes the shuffling . we now show a shuffling weight function always exists by giving an algorithm which will generate one . * set @xmath49 and construct a @xmath50 array where the element in the @xmath51th entry of the array ( @xmath52 and @xmath53 is @xmath54 , where @xmath55 is the remainder when @xmath56 is divided by @xmath57 . * while there are still cells in the array which have not been crossed off do the following : * * construct a directed graph on the vertices @xmath58 by letting @xmath59 where @xmath60 is the column and @xmath61 is the @xmath62 entry in the _ lowest _ cell of column @xmath60 which has not been crossed out . ( if all the cells in column @xmath60 have been crossed out then @xmath60 will be an isolated node . ) * * find all directed cycles in the graph that was constructed . for each edge in the directed cycle let @xmath63 , where @xmath64 comes from the cell that generated the edge . * * cross out all cells which were used to assign a weight . * * increase @xmath65 by @xmath66 . an example of the algorithm being carried out when @xmath67 and @xmath68 is shown in the appendix . ( the reader is encouraged to try this algorithm out to generate the weight function for @xmath17 and @xmath16 given above and for @xmath69 and @xmath70 given in section [ sec : conclusion ] . ) the above algorithm generates a shuffling weight function . first let us observe that the algorithm will terminate so @xmath45 is a well defined function . this follows since each column is referred to exactly @xmath2 times and so if all of the entries in a column have been crossed out no edge will be directed towards it . in particular , if we ignore isolated vertices then the directed graph generated in step 2i will have all vertices with outdegree @xmath71 and so it must contain at least one directed cycle ( in fact one directed cycle for each connected component ) . therefore at every stage we will continue to cross out cells unless all cells have been crossed out so the algorithm will terminate . the first condition of a shuffling weight function ( that weights in the blocks match weights in the columns ) is satisfied because we are pulling out cycles . namely , if we look at a cell @xmath72 in the @xmath60th column , this can be understood as the subscript @xmath64 will be placed into the @xmath62th column ( when shuffling ) . what we need is to make sure there is some subscript which will go into the @xmath60th column that also will be assigned the same weight . but this happens because we assign the entire directed cycle the same weight and by step 2ii the vertex preceding @xmath60 will give a subscript that maps into the @xmath62th column . the second condition of a shuffling weight function ( that weights in each block are increasing ) is easily satisfied since the blocks form the columns of the array generated in step 1 , and in columns the weights assigned to the cells are increasing ( since in each round at most one cell in each column will be assigned the current weight . ) the algorithm used the lowest cell which had not been crossed out . we could also have used the highest cell which was not crossed out and then decreased the weight by @xmath71 . this will essentially generate the same picture ( where we rotated the arrays by @xmath73 and let @xmath74 in the entries and in the directed graphs generated in step 2i of the algorithm . another important thing to note is the directed cycles we are pulling out are the same directed cycles ( but with the @xmath62 terms replaced by the @xmath64 terms from the corresponding cells ) found in the shuffling poset . if we combine this with the previous idea of going from bottom to top we get the following observation . in the shuffling poset @xmath26 if and only if @xmath75 . something else worth noting is that if we look at the shuffling weight function produced in the appendix we see there is some symmetry , i.e. , @xmath76 is independent of @xmath46 . so for example we have @xmath77 . when this happens we will call this a _ symmetrical _ weight function . the shuffling weight function produced by the algorithm is symmetric . it is easy to show a symmetric weight function always exists for every @xmath24 and @xmath2 . for instance let @xmath78 be the weight function generated by the given algorithm ( where we use the lowest cells and work our way up ) and let @xmath79 be the weight function generated by the modified algorithm ( where we use the highest cells and work our way down ) . then the weight function @xmath80 is still a shuffling weight function and by the symmetry of the two algorithms will be symmetric . in the preceding section we saw that a weight function could be used to give us the shuffling poset , which in turn gives us another representation of how to shuffle . one natural question to ask is what can we say about the lengths of the directed cycles in the shuffling poset . in this section we will show in the special case when @xmath0 and @xmath3 this question has an easy answer . note this will cover all the cases when @xmath2 is a prime ( in particular @xmath81 ) . first , for @xmath3 we observe one shuffling weight function can be found using the base @xmath2 expansion of subscripts . [ lem : qk ] let @xmath82 with @xmath3 and let @xmath83 be the base @xmath2 expansion of @xmath84 . then @xmath85 is a shuffling weight function . let @xmath33 and @xmath86 be the base @xmath2 expansion of @xmath87 , and let the base @xmath2 expansion of @xmath7 be @xmath88 where @xmath89 is in the @xmath90th slot and @xmath91 ( this follows from @xmath25 ) . in particular , for @xmath92 the base @xmath2 expansion of @xmath93 will be @xmath94 which has @xmath95 . on the other hand the base @xmath2 expansion of @xmath96 will be @xmath97 which has @xmath98 ( where @xmath99 is the remainder of @xmath100 when divided by @xmath101 ) . because @xmath102 then @xmath103 will cycle through all @xmath2 possibilities as @xmath46 goes from @xmath104 to @xmath105 . it follows the first condition of a shuffling weight function is satisfied . the second condition of a shuffling weight function is satisfied since the base @xmath2 expansion of @xmath93 for @xmath92 is @xmath106 and so @xmath107 from this it follows @xmath35 . one nice feature about this case is the the rule for mapping subscripts is easy to describe . namely , if the base @xmath2 expansion of @xmath84 is @xmath108 then the map is @xmath109 ( there are two things to check , one is that both these terms have the same weight and the other is that the column that contains @xmath84 will map to the block that contains @xmath110 , both conditions are easily checked . ) we will now use this map to determine the cycle lengths in the shuffling poset . [ thm : period ] let @xmath0 with @xmath111 , and let @xmath112 denote the multiplicative order of @xmath2 modulo @xmath100 . then the length of a cycle in the shuffling poset when we divide @xmath24 into @xmath2 equal stacks for shuffling is a divisor of @xmath4 . further , there is a cycle of length @xmath4 in the shuffling poset . before we begin we note by our assumption that @xmath113 and so the multiplicative order is well defined . as a check we note @xmath114 since @xmath115 , this agrees with the diagram given above for @xmath17 and @xmath16 . consider a cycle starting at @xmath116 , and suppose that the base @xmath2 expansion of @xmath116 is @xmath117 . using we have after @xmath101 steps we will be at @xmath118 in particular , after we have taken @xmath101 steps , the last @xmath101 terms in the base @xmath2 expansion will agree with the last @xmath101 terms in the base @xmath2 expansion of @xmath116 . now suppose we repeat this @xmath119 times ( so a total of @xmath120 steps ) . then we have @xmath121 for some @xmath119 we will be back where we started if @xmath122 multiplying both sides by @xmath123 and simplifying this is equivalent to @xmath124 looking at the base @xmath2 expansion of @xmath116 we have @xmath125 for some @xmath126 , if we now substitute this in and simplify we get @xmath127 or @xmath128 since @xmath129 , if @xmath130 then @xmath131 and this condition is satisfied . in particular , after taking @xmath4 steps then @xmath132 for each value of @xmath116 , and so each cycle must be some divisor of @xmath4 . we are implicitly using @xmath133 . to see why this is true , let @xmath134 then @xmath135 , and so @xmath136 but we also have in general @xmath137 combining these two statements we have @xmath138 showing @xmath101 is a divisor of @xmath134 . finally , to show there is a cycle of length @xmath4 we note also holds for @xmath139 , which corresponds to @xmath140 and @xmath141 . so reduces to finding the smallest value of @xmath120 so @xmath142 which is clearly @xmath130 . the preceding weight function relied on having @xmath3 ( this was used in the lemma to make sure we hit all of the residue classes modulo @xmath2 , and then in the theorem [ thm : period ] by giving a simple rule for mapping ) . we now present a weight function that works in more cases ( but from the definition it will agree with the previous weight function when @xmath3 ) . for example , the following weight function will work for the case when @xmath2 is any prime power . let @xmath82 with @xmath143 and let @xmath144 be the base @xmath2 expansion of @xmath84 . then @xmath145 is a shuffling weight function . for @xmath28 let @xmath146 be the base @xmath2 expansion of @xmath87 . then for @xmath92 , the base @xmath2 expansion of @xmath93 is @xmath147 and so @xmath148 using this , we note the second condition for a shuffling weight function is easily satisfied . since @xmath25 , the base @xmath2 expansion of @xmath7 is @xmath149 . also we have @xmath150 , and so @xmath151 . so for @xmath92 the base @xmath2 expansion of @xmath152 is @xmath153 , i.e. , it agrees in the first @xmath154 slots with the expansion of @xmath87 , so we need to understand @xmath155 . for a given value of @xmath46 we have @xmath156 . since @xmath157 is relatively prime to @xmath2 , this takes on @xmath158 different values that differ by a multiple of @xmath159 from @xmath160 ( modulo @xmath2 ) , and we will attain each one of these values @xmath159 times as @xmath161 ranges over its possible values . in particular we have , @xmath162 fix a possible value for @xmath163 and let @xmath164 be the first value of @xmath46 that gives us @xmath163 , and let @xmath165 be the value of @xmath166 for this corresponding @xmath46 . in particular , the only values of @xmath46 where we will be at the fixed value of @xmath163 are @xmath167 for @xmath168 . for these values of @xmath161 we have @xmath169 ( the last step follows from noting @xmath170 and that when we divide this by @xmath159 all but the first two terms will have at least one factor of @xmath2 . ) so we have @xmath171 since @xmath143 then this covers all of the residue classes modulo @xmath159 . combined with what we know about @xmath163 then we have @xmath172 so we have @xmath173 where @xmath47 which ranges over @xmath174 as @xmath46 ranges over @xmath174 . the problem is even though we have a simple weight function , the rule for mapping is not simple , and so we have no similar result as theorem [ thm : period ] . we note one of the results coming out of the proof of theorem [ thm : period ] is when @xmath3 the cycle in the shuffle poset which contains @xmath71 has maximal length . this no longer needs to hold when @xmath175 . if we look at the example shown in the appendix , we can see the cycle which will contain @xmath71 has length @xmath176 but the maximal length of a cycle is @xmath177 . in section [ sec : cycles ] we saw a way to represent our shuffling in terms of a poset of cycles . we will now exploit this poset to find stacks of cards which are fixed or periodic under shuffling . to find the fixed stacks , we observe when looking at the cycles of the shuffling poset in a stack which is fixed under shuffling there are two necessary conditions : * if there is an edge between levels in the shuffling poset then the label on the lower level must be at least as great as the label on the higher level . ( otherwise we would swap labels and we would not be fixed . ) * all the labels in a cycle on a level in the shuffling poset must agree . ( otherwise when we shift by one in the cycle the stack is not fixed . ) it is easy to see that these conditions are also sufficient . we now form a poset , which we call the _ fixed poset _ , based off the shuffling poset . namely , each cycle goes to one element in the poset ( at the same height as before ) and we connect an edge between two corresponding cycles if one cycle contains @xmath178 while the other contains @xmath179 from the same column and there is no @xmath180 in the column with @xmath181 . ( technically , by condition ( i ) above we would want to connect all cycles which are connected by an edge in the shuffle poset , but by transitivity we only need to consider edges which can not be broken down further . ) some examples of fixed posets are given in figure [ fig : fixpo ] . using the two conditions , the fixed stacks are now found by assigning labeled cards to each element in the fixed poset so the labels are weakly increasing with respect to the fixed poset . this allows us to quickly and easily find fixed stacks for a given @xmath24 and @xmath2 . it also helps us to answer whether or not fixed stacks with some given property can exist . for example if @xmath182 and @xmath81 it is easy to see from the fixed poset there is no fixed stack with four labels each label with 13 cards ( i.e. , such as in a standard deck of cards with the labels being the suits in some order ) . note the fixed posets in figure [ fig : fixpo ] are symmetric in that if we flipped it upside down we would have the same poset . this will always happen when our weight function is symmetric ( and as already noted we can always find such a weight function ) . one question to consider is the number of fixed stacks . this is dependent on both the fixed poset and the number of labels . as an example , for a fixed @xmath2 and a fixed order of the labeling we can have dramatically different behavior for various values of @xmath24 . for example , suppose we have two labels @xmath183 and @xmath81 . if @xmath184 , then by theorem [ thm : period ] all the cycles have length at most @xmath101 and there are @xmath185 levels in the fixed poset using the weight function given in lemma [ lem : qk ] . we can then assign all of the cycles with height @xmath186 in the poset to have label @xmath71 and then take any collection of posets from the layer at height @xmath187 to have label @xmath71 . there are at least @xmath188 cycles on the @xmath187 level for some constant @xmath189 , and since we can take any subset of them and form a fixed stack then there are at least @xmath190 fixed stacks for @xmath184 ( in particular this is super - polynomial ) . by comparison if we consider @xmath191 , then it is easy to show it has a fixed poset of the following form . so in particular there are exactly @xmath192 fixed stacks for @xmath191 . to make the comparison more concrete we have for @xmath193 there are @xmath194 fixed stacks , while for @xmath195 there are 128 fixed stacks . when looking for periodic stacks we will again set up a poset , but instead of having cycles as elements in the poset we will have individual subscripts be the elements the basic idea is to consider what can keep a card at a given subscript from dropping down as we go through repeated shufflings ( because we are periodic the only operation that will happen as we shuffle is shifting the cycles ) . in particular , an element @xmath84 can only drop down if for some @xmath196 and some @xmath197 we have the following edges in our shuffling poset : @xmath198 in this case we need to make sure the card in @xmath196 will not cause the card in @xmath84 to sink . to do this we draw the poset where the elements are the subscripts ( as before we can place @xmath84 at level @xmath199 ) , and we connect an edge between @xmath84 and @xmath196 if we have the edges in our shuffling poset as indicated above _ and _ there is no @xmath180 in the same column as @xmath200 and @xmath201 so @xmath202 . we will call this the _ periodic poset_. an example of the situation is shown in figure [ fig:1624 ] . as before , the periodic stacks are now found by assigning labeled cards to each subscript in the periodic poset so the labels are weakly increasing with respect to the periodic poset . this allows us to quickly and easily find periodic stacks for given @xmath24 and @xmath2 . the possible periods of the periodic stacks are determined by the size of the cycles in the shuffling poset . namely , the possible periods are the divisors of the least common multiple of the cycle lengths in the shuffling poset . in the case when @xmath3 then theorem [ thm : period ] shows the possible periods of stacks are divisors of @xmath4 , on the other hand it is easy to construct a periodic stack for any period dividing @xmath4 . we have seen how to find a weight function which can in turn allow us to represent our shuffling in a shuffling poset . this poset can then be modified to help us find fixed stacks , periodic stacks , and also tell us which periods are possible . one of the problems we have not addressed is how quickly a stack will settle into a periodic stack . as with the number of fixed stacks this depends highly on @xmath24 . for example , it is not hard to see ( i.e. , using the periodic poset ) that it takes no more than @xmath203 shuffles to settle into a periodic orbit ( where @xmath126 is the least common multiple of the cycle lengths and @xmath185 the number of levels in the shuffling poset ) . so for example when @xmath184 and @xmath81 then we need at most @xmath204 steps . on the other hand for @xmath205 it is easy to construct a stack that takes exactly @xmath206 steps to get into a periodic stack . there are still many questions that remain . one of the biggest problems is trying to understand how the weight function works for arbitrary @xmath24 and @xmath2 . for instance the weight functions given in section [ sec : gcd ] do not apply for the case @xmath69 and @xmath70 . in this case the algorithm for finding a weight function generates the following : @xmath207 \begin{array}{\|\|c\|\|c\|c\|c\|c\|c\|c\|c\|c\|c\|c\|c\|c\|\| } \hline\hline n&12&13&14&15&16&17&18&19&20&21&22&23\\ \hline \varphi(n)&1&2&3&4&7&8&2&3&4&5&8&9 \\ \hline\hline \end{array } \end{array}\ ] ] one notable difference between this weight function and the weight function when @xmath3 or @xmath143 is that in the blocks the weight function does not consist of consecutive numbers , i.e. , it has gaps . determining why these gaps are there and where they will appear given @xmath24 and @xmath2 will go a long way to understanding the shuffling weight function . another important question in regards to the shuffling posets is understanding the possible cycle lengths . we understand what is going on for the case when @xmath3 , but all other cases remain open . for example , is it true the least common multiple of the cycle lengths is the length of the longest cycle ? we can also consider what happens when instead of only considering a single type of shuffling we consider combining the @xmath208 different shuffling rules that come from all the possible rearrangements of the ordering of the labels . and of course , perhaps the most important thing missing right now is a good magic trick that can be performed using this shuffling rule , which was the original motivation of larry carter and j .- c . reyes who first suggested this problem ! 99 persi diaconis , ron graham and william kantor , _ the mathematics of perfect shuffles _ , adv . in appl * 4 * ( 1983 ) , 175196 . s. brent morris , _ magic tricks , card shuffling and dynamic computer memories _ , maa spectrum , mathematical association of america , washington , d.c . , ( 1998 ) xviii + 148 pp . we implement the algorithm given in section [ algo ] to find the weight function in the case @xmath67 and @xmath68 . the steps are shown below . the generated weight function is given in the following table . @xmath209 \begin{array}{\|\|c\|\|c\|c\|c\|c\|c\|c\|c\|c\|c\|c\|c\|c\|c\|c\|c\|\| } \hline\hline n&17&18&19&20&21&22&23&24&25&26&27&28&29&30&31\\ \hline \varphi(n)&2&3&4&2&3&4&5&3&4&5&6&4&5&6&7 \\ \hline\hline \end{array } \end{array}\ ] ]	we consider a problem of shuffling a deck of cards with ordered labels . namely we split the deck of @xmath0 cards ( where @xmath1 is maximal ) into @xmath2 equally sized stacks and then take the top card off of each stack and sort them by the order of their labels and add them to the shuffled stack . we show how to find stacks of cards invariant and periodic under the shuffling . we also show when @xmath3 the possible periods of this shuffling are all divisors of @xmath4 .
the production of ultra - cold neutrons ( ucn ) @xcite in superfluid helium and their subsequent storage allowing the build up of densities much greater than would be achievable at thermal equilibrium with the helium ( superthermal ucn source)@xcite is the foundation for several current and proposed experiments with ucn @xcite . although calculations of the production rate have been presented previously @xcite the early calculations concentrated on the production due to the resonant ( 8.9 ) phonons . measurements @xcite have shown a rather narrow peak in the production cross section as expected . in the course of the first calculation @xcite estimates had been made of the contribution due to higher energy incident neutrons downscattering on the multiphonon excitation tail in @xmath0 and these neutrons were found to contribute about the same amount of ucn production as the resonant neutrons for a reasonable assumed spectrum . however as the initial work was part of a proposal for an experiment it was felt that it would be more conservative not to include this in the initial estimates of ucn density i.e. to keep this in hand as a factor of safety for the predicted intensity for the new source . however the fact that multiphonon processes would increase the ucn production rate was mentioned . in recent years the comparison between theory and experiment has improved significantly and there are several proposals to operate a superthermal helium ucn source with a monochromatic beam of 8.9 neutrons , so the question of the production rate due to multiphonon processes has some practical importance . in addition , due to a vigorous experimental program @xcite , much more detailed information concerning @xmath1 has become available . for these reasons it has become necessary to re - examine the situation . w. schott , in particular , has emphasized the importance of the ucn production due to higher energy neutrons . this work is concerned with calculations of the ucn production rate when a volume of superfluid he@xmath2 at temperatures of @xmath3 is exposed to a neutron flux . the steady state ucn density in the source is given by@xmath4 where @xmath5 is the production rate , @xmath6 and @xmath7 is the ucn lifetime against all losses in the helium filled storage chamber . lifetimes @xmath8 seconds should be achievable . the storage time in superfluid helium has been measured in @xcite . a neutron at rest can absorb energy @xmath9 and momentum @xmath10 with @xmath11 with ( @xmath12 mev / @xmath13 ) and contrarily a neutron with this energy and momentum can come to rest after transferring its energy and momentum to the superfluid . for single phonon interactions , which are usually dominant , the superfluid can only exchange quantities of energy and momentum that are related by the dispersion curve@xmath14 where the last relation is an approximation to simplify the discussion . then neutrons can only come to rest by emission of a single phonon if they have the resonant energy @xmath15 given by the intersection of equations ( [ one ] ) and ( [ two])@xmath16 in addition to one phonon interactions there are multiple phonon interactions occurring at higher input energies that can also produce ucn . these are the main subject of this work . ucn production in helium can be calculated as follows : the differential cross section for neutron scattering is given by the fourier transform of the van hove correlation function @xmath17 @xcite : @xmath18 and @xmath17 has been measured in great detail @xcite . @xmath19then @xmath20 since the limits on @xmath21 are : @xmath22 @xmath23 @xmath24 ( @xmath25 is the ucn momentum ) we have @xmath26 . then @xmath27 where we assumed @xmath17 is constant over the narrow range @xmath28 . @xmath29 the ucn production rate is given by @xmath30 de_{u}%\ ] ] then @xmath31 \int_{0}^{k_{c}}k_{u}^{2}dk_{u}\nonumber\\ & = n_{he}4\pi b^{2}\alpha^{2}\left [ \int\frac{d\phi\left ( k_{1}\right ) } { de}s(k_{1},\omega=\frac{\alpha k_{1}^{2}}{2})dk_{1}\right ] \frac{k_{c}^{3}% } { 3}\quad ucn / cm^{3}/\sec\label{111}%\end{aligned}\ ] ] where @xmath32 , @xmath33 are the critical ucn energy and wave vector of the walls of the storage chamber and @xmath34 is the energy spectrum of the incoming flux . note @xmath35 is in @xmath36 the one phonon production rate @xcite is found by evaluating the integral in equ.([111 ] ) over the one phonon peak . ( note that @xmath37 @xmath38 ) . @xmath39 where @xmath40 @xcite and @xmath41 @xcite with @xmath42 the neutron velocity at the critical energy for ucn production , @xmath15 , and @xmath43 the phonon group velocity at the critical energy . we have replaced @xmath44 where the subscript means the integration is carried out across the phonon peak with @xmath45 where @xmath46 is the contribution to the structure factor from the one phonon scattering @xcite and @xmath47 represents the increase in the integral due to the fact that the two curves ( [ one ] ) and ( [ two ] ) cross at an angle , so the path of integration is longer than in the calculation of ( [ s1 ] ) , as was first pointed out by pendlebury@xcite . numerical integration of the measured @xmath48 @xcite across the phonon peak , ( fig . 1 ) confirms this procedure . ( note that the factor of 2 in equation 9 of [ ] , which holds in the case of a linear dispersion relation should be replaced by @xmath49 for the real , non - linear dispersion relation ) thus ( see equation [ 100 ] ) @xmath50 where @xmath51 is in units of @xmath52 . [ fig1 ] shows @xmath53 as a function of @xmath21 obtained from [ ] . we have extrapolated the data above @xmath54@xmath38 in fig . [ fig1 ] , taking into account the known value of @xmath55 for the one phonon and multiphonon contributions @xcite . measurements @xcite show that @xmath56 is essentially zero for @xmath57@xmath38 . as an illustration we apply the above technique to various measured and calculated slow neutron spectra that are under discussion as locations for various superthermal ucn source projects . we will take @xmath58@xmath38 corresponding to the critical energy for beryllium . [ ptb ] figure1.ps north carolina state university researchers have proposed to locate a ucn source in the thermal column of the campus 1-mw pulstar reactor after removing the graphite . the expected average cold neutron flux in the ucn converter was estimated using mcnp simulations with a detailed model of the pulstar core and a conceptual model of the ucn source . the core s 4 control blades and 25 fuel assemblies , each consisting of a 5 x 5 array of fuel pins containing 4 % enriched uo2 , were faithfully represented . the model of the ucn source consisted of a cold neutron source surrounding a ucn converter about 10-cm deep in a tank of d2o in the thermal column void . the cold source moderator was 1-cm - thick cup - shaped solid methane ( 408 g , 22 k ) contained in a 2-mm - thick - wall aluminum chamber ( 1164 g ) . the ucn converter was modeled by a disk of liquid ortho - deuterium ( 4-cm thick x 14-cm diameter , 20 k ) in a 5-mm - thick - wall aluminum container ( 484 g ) . neutrons leaving a bare face of the reactor core were channeled into the d@xmath590 tank by a 44-cm diameter x 75-cm long void in a beryllium assembly located between the reactor core and the reactor tank wall . the simulated spectrum is shown in fig . [ fig1 ] , along with the spectrum of some other sources discussed below and @xmath60 . for this spectrum @xmath61 = 5.22\times10^{9}$ ] substituting into equation ( [ 111 ] ) : @xmath62 we have@xmath63 for the one phonon contribution we have , taking @xmath64 from fig . [ fig1 ] and using equation ( [ 101 ] ) @xmath65 these figures will be reduced by a factor of 2 when the flux ( calculated here for a modest source volume as described above ) is averaged over a more reasonable 20 liter volume . the effects of construction materials in a realistic design have not yet been taken into account . the advantages of such a design were discussed in [ ] . the main point is that the ucn production process is independent of the direction of the incident neutrons so that use is made of the full solid angle of the flux . installations at a beam position , as discussed in the next section , lose a solid angle factor of about 10@xmath2 [ ] in comparison to installations inside a 4@xmath66 flux , although such sources outperform any other type of ucn source that could be installed at the same beam position . it has been proposed to place a superthermal source of ucn on a monochromatic 8.9 beam at a guide tube at the sns.@xcite it is interesting to ask what the loss of ucn production rate associated with the use of a monochromatic beam would be in the case of the simulated spectra available for the planned guide tubes . spectra have been simulated for two types of guide : an ordinary supermirror guide and a so - called ballistic guide using tapered sections to focus and defocus the beam at the borders of a long free - flight region . this is an interesting case for the present discussion as the ballistic guide has more flux at the critical 8.9 wavelength but less flux at shorter wavelengths ( @xmath67 ) . the predicted fluxes for the two types of guide are shown in fig . [ fig2 ] , while fig . [ fig1 ] shows a comparison of the sns ballistic guide with the other sources considered in this paper . [ ptb ] figure2.eps the result for the ordinary guide is that @xmath68 so that @xmath69 for the one phonon production rate we find using ( [ 101])@xmath70 for the ballistic guide we have @xmath71 so that @xmath72 and the one phonon production rate is : @xmath73 . thus we see that the losses associated with rejecting the higher energy neutrons in the beam by the monochromator are less significant than the losses ( 50% ) @xcite connected with the monochromator and associated beam bender . this spectrum was measured at the nl4 neutron guide at the hahn meitner institut , berlin by time of flight . in this context we should remind the reader that the measurement of the spectrum from a neutron guide is quite complex as the spectra in general are very strong functions of the flight direction and position in the beam cross section . for the measured spectrum ( normalized to 1 at the peak of @xmath74 ) we find:@xmath75 = .025\ ] ] yielding@xmath76 the one phonon contribution for this spectrum is found ( equ . [ 101 ] ) as @xmath77 [ ptb ] figure3.ps fig . [ fig3 ] shows the spectrum and calculated ucn production rate as a function of the wavelength of the incident neutrons , @xmath78 , similar to what would be observed in a time of flight measurement . the results are summarized in the table below . the column labelled maxwell refers to a maxwellian spectrum corrected for an ideal guide transmission @xmath79 and cut off at @xmath80 minimum wavelength . we see that the multiphonon contribution as expected is a rather strong function of the source spectrum varying from less than to slightly more than the one phonon production rate for the realistic spectra considered here . [ c]\|l\|l\|l\|l\|l\|l\| & nc state@xmath81 & sns ord@xmath81 & sns ball@xmath81 & hmi a.u . & maxwell + multi - ph & 490 & 1.0 & .94 & 4.7 & 1.7 + single ph & 375 & 1.8 & 2.4 & 5.5 & 1.5 + mph/1ph & 1.4 & .55 & .4 & .85 & @xmath821.13 + thus we see that the inclusion of the multiphonon production amounts to at most a little more than a factor of 2 increase in ucn production . in the case of a cold beam the multiphonon contribution is a small correction to the single phonon production so that the use of a monochromatic beam , which offers significant operating advantages @xcite , would be accompanied by a minor loss in ucn production . in contrast , for sources where the helium is exposed to the total thermal flux as in the north carolina state proposal or at a dedicated spallation source , the multiphonon contribution can amount to slightly more than a factor of 2 increase in ucn production .	ultra - cold neutrons ( ucn ) are neutrons with energies so low they can be stored in material bottles and magnetic traps . they have been used to provide the currently most accurate experiments on the neutron life time and electric dipole moment . ucn can be produced in superfluid helium at significantly higher densities than by other methods . the predominant production process is usually by one phonon emission which can only occur at a single incident neutron energy because of momentum and energy conservation . however ucn can also be produced by multiphonon processes . it is the purpose of this work to examine this multiphonon production of ucn . we look at several different incident neutron spectra , including cases where the multiphonon production is significant , and see how the relative importance of multiphonon production is influenced by the incident spectrum .
consider a nonequilibrium stochastic system in thermal contact with a heat bath . it is driven out of equilibrium by a time - dependent perturbation or a nonconservative force . due to absence of the detailed balance , a nonequilibrium system is characterized by a nonzero probability current and a net energy flow . recently , nonequilibrium fluctuations of thermodynamic quantities such as a work , an entropy production , and a heat have been attracting growing interests . various types of fluctuation theorems governing the nonequilibrium fluctuations have been found in a thermostated shearing fluid @xcite , systems driven by a nonequilibrium work @xcite , stochastic langevin systems @xcite , master equation systems @xcite , systems in a nonequilibrium steady state @xcite , general stochastic systems @xcite , feedback control systems @xcite , systems with odd parity variables @xcite , and so on . theoretically , the fluctuation theorems allow us to study the time irreversibility of nonequilibrium systems . at the same time , they play an important role in studying small - sized systems such as colloidal particles @xcite , bio - molecules @xcite , and molecular motors @xcite , where thermal fluctuation effects are important . further theoretical and experimental studies are found in recent review papers @xcite . we are interested in mutual correlations among the thermodynamic quantities . most studies have been focused on the fluctuations of an individual thermodynamic quantity with a few exceptions @xcite . the current study is motivated by our recent work on the modified fluctuation relation for the heat @xcite . the amounts of a work @xmath0 and a heat @xmath1 during a nonequilibrium process over a time interval @xmath2 are constrained by the thermodynamic first law @xmath3 where @xmath4 denotes the change in the internal energy of the system . with nonzero net energy flow , both @xmath0 and @xmath1 scale with the time interval @xmath2 while their difference @xmath4 does not on average . hence , one may expect that the heat would follow the same fluctuation theorem obeyed by the work in the large @xmath2 limit . however , various studies found that the fluctuation theorem breaks down for the heat @xcite . the breakdown suggests that the effect of the the boundary term @xmath4 may persist even in the infinite time limit @xcite . moreover , the modified fluctuation relation for the heat derived in ref . @xcite suggests that the correlation between the heat and the energy change plays an important role in nonequilibrium processes . based on this motivation , we investigate the mutual correlations among the thermodynamic quantities in a linear diffusion system . it is simple , but exhibits various nontrivial genuine nonequilibrium phenomena @xcite . we develop a path integral formalism to study the joint probability distribution analytically . our study gives a hint on the reason why the fluctuations of the work and the heat are different . this paper is organized as follows : in sec . [ sec2 ] , we introduce an overdamped langevin dynamics for a nonequilibrium system driven by both a time - dependent perturbation and a nonconservative force , and review the stochastic thermodynamics formalism for the fluctuation theorem . in sec . [ sec3 ] , we derive the modified fluctuation relation for the heat in a general setting . in ref . @xcite , we only considered a time - independent nonconservative force . we extend the formalism to include a time - dependent driving force . in sec . [ sec4 ] , we investigate the fluctuations and the correlations in a two - dimensional linear diffusion system in detail . we develop an analytic method in [ appendix ] , which allows us to calculate the joint probability distributions of the thermodynamic quantities analytically . we find that there is a strong negative correlation between @xmath1 and @xmath4 . we conclude the paper with summary in sec . consider an overdamped dynamics of a nonequilibrium system in thermal contact with a heat bath at temperature @xmath5 . when there are @xmath6 degrees of freedom , a configuration is described by an @xmath6-dimensional column vector @xmath7 . the langevin equation reads as @xmath8 where @xmath9 is a force and @xmath10 is a thermal noise satisfying @xmath11 with the inverse temperature @xmath12 . the superscript @xmath13 denotes the transpose . the damping coefficient and the boltzmann constant are set to unity . in general , the force @xmath14 is a sum of conservative and nonconservative forces as @xmath15 with an energy function @xmath16 . the energy function may depend on a time - dependent external parameter @xmath17 , called a protocol . the nonconservative force @xmath18 can not be written as a gradient of any scalar function . when @xmath19 and @xmath20 is independent of @xmath21 , the system relaxes into a thermal equilibrium state characterized by the boltzmann distribution @xcite @xmath22\ ] ] with the free energy given by @xmath23 .\ ] ] a nonzero @xmath18 or a time - dependent protocol @xmath24 drives the system out of equilibrium . during the time evolution , the system exchanges the energy with the environment . suppose that the system evolves in time along a stochastic path or trajectory @xmath25 during the time interval @xmath26 . then , the nonequilibrium work done on the system is given by @xcite @xmath27 = \int_{t_i}^{t_f } dt \left [ \frac{d\alpha}{dt } \frac{\partial v({{\bm{x}}},\alpha)}{\partial \alpha } + \dot{{\bm{x}}}^t \cdot { { \bm{f}}}_{nc}({{\bm{x } } } ) \right ] , \ ] ] where the first term is the thermodynamic work required to change the external parameter and the second term is the work done by the nonconservative force . the change in the internal energy is given by @xmath28 = v({{\bm{x}}}(t_f),\alpha(t_f ) ) - v({{\bm{x}}}(t_i),\alpha(t_i ) ) .\ ] ] it can be rewritten as @xmath29 & = & \int_{t_i}^{t_f } dt \ \frac{d}{dt}v({{\bm{x}}}(t),\alpha(t ) ) \nonumber \\ & = & \int_{t_i}^{t_f } dt \ \left [ \frac{d\alpha}{dt}\frac{\partial v}{\partial \alpha } + \dot{{\bm{x}}}^t \cdot \bm\nabla_{{\bm{x}}}v\right ] . \label{e_def}\end{aligned}\ ] ] the heat dissipated into the bath is then given by @xmath30 = { { \mathcal{w}}}[{{\bm{x}}}(t ) ] - \delta { { \mathcal{e}}}[{{\bm{x}}}(t ) ] = \int_{t_i}^{t_f } dt\ \dot{{\bm{x}}}^t \cdot { { \bm{f}}}.\ ] ] note that the integral in ( [ w_def ] ) , ( [ e_def ] ) , and ( [ q_def ] ) is the stratonovich integral @xcite . since the system follows the stochastic langevin dynamics , thermodynamic quantities @xmath0 , @xmath31 , and @xmath1 are random variables . stochastic thermodynamics predicts several interesting identities , known as the fluctuation theorems , for their probability density functions ( pdfs ) . we present a brief summary on the fluctuation theorems @xcite . the key feature behind the fluctuation theorem is that the thermodynamic quantity can be written in terms of a relative entropy @xcite between the path probabilities in forward ( f ) and reverse ( r ) processes . the f process is specified by a protocol @xmath24 , a nonconservative force @xmath18 , and a pdf @xmath32 for the initial configuration . the r process , corresponding to the f process , is specified by the time - reversed protocol @xmath33 , the same nonconservative force @xmath18 , and a pdf @xmath34 for the initial configuration . consider a following quantity associated with a path @xmath35 @xmath36 \equiv \ln \left ( \frac{{{\mathcal{p}}}[{{\bm{x}}}(t)]}{{{\mathcal{p}}}^r[{{\bm{x}}}^r(t ) ] } \right ) \ , \ ] ] where @xmath37 $ ] and @xmath38 $ ] are the probability distribution of the system following a trajectory @xmath25 in the f and r processes , respectively , and @xmath39 is the time - reversed trajectory . such a quantity satisfies the identity @xmath40 } \right\rangle = 1 \ , \ ] ] where the average is taken over all paths @xmath25 in the f process . the jensen s inequality then yields that @xmath41 \rangle \geq 0 \ .\ ] ] the path probability distribution is given by @xmath42 = { { \mathcal{t}}}^{(r)}[{{\bm{x}}}(t)\|{{\bm{x}}}(t_i ) ] p^{(r)}_i({{\bm{x}}}_i)\ ] ] with the conditional probability @xmath43 $ ] for a path @xmath25 starting at @xmath44 in the f ( without superscript ) and r ( with superscript @xmath45 ) processes . using the onsager - machlup formalism @xcite , one can show that the ratio between the conditional probabilities is equal to the heat dissipation @xcite : @xmath46 = \beta^{-1 } \ln\left ( \frac{{{\mathcal{t}}}[{{\bm{x}}}(t)\|{{\bm{x}}}_i ] } { { { \mathcal{t}}}^r[{{\bm{x}}}^r(t)\|{{\bm{x}}}_f]}\right)\ ] ] with @xmath47 and @xmath48 . through this relation , @xmath49 can represent a thermodynamic quantity with a suitable choice of @xmath32 and @xmath34 . such a quantity then automatically satisfies the identity of ( [ ift ] ) , which is referred to as an integral fluctuation theorem ( ift ) . one can obtain the well - known itf for the total entropy change @xmath50 + \beta{{\mathcal{q}}}\ ] ] by choosing @xmath51 as the pdf @xmath52 of finding the system , evolving from the initial pdf @xmath53 at @xmath54 in the f process , in configuration @xmath55 at time @xmath56 . the first term corresponds to the change in the shannon entropy of the system @xmath57 , while the second term the change of the heat bath entropy . one can also obtain the itf for the nonequilibrium work , known as the jarzynski equality @xcite , by choosing @xmath58 and @xmath59 with the boltzmann distribution in ( [ p_eq ] ) . with this choice , one finds that @xmath60 with the free energy difference @xmath61 . in a certain circumstance , the functional @xmath62 $ ] and the functional @xmath63 \equiv \ln[{{\mathcal{p}}}^r[{{\bm{x}}}(t)]/{{\mathcal{p}}}^{rr}[{{\bm{x}}}^r(t)]]$ ] may satisfy a relation @xmath64=-{{\mathcal{y}}}[{{\bm{x}}}(t)]$ ] . it holds only when the reverse process of the r process is equivalent to the original f process , i.e. , @xmath65={{\mathcal{p}}}[{{\bm{x}}}(t)]$ ] , which is called an involution property @xcite . then , the pdf @xmath66-y)\rangle$ ] for the f process and the pdf @xmath67-y)\rangle_r$ ] for the r process satisfy the identity @xcite @xmath68 which is called a detailed fluctuation theorem ( dft ) in comparison to the ift . the dft implies the ift , while the converse is not true in general . using ( [ y_def ] ) and the involution property , one can derive ( [ dft ] ) as follows : @xmath69 \delta({{\mathcal{y}}}[{{\bm{x}}}]-y ) { { \mathcal{p}}}[{{\bm{x } } } ] \\ & = & \int[{d}{{\bm{x } } } ] \delta({{\mathcal{y}}}[{{\bm{x}}}]-y ) { { \mathcal{p}}}^r[{{\bm{x}}}^r ] e^{{{\mathcal{y}}}[{{\bm{x } } } ] } \\ & = & \int[{d}{{\bm{x}}}^r ] \delta(-{{\mathcal{y}}}^r[{{\bm{x}}}^r]-y ) { { \mathcal{p}}}^r[{{\bm{x}}}^r ] e^{-{{\mathcal{y}}}^r[{{\bm{x}}}^r ] } \\ & = & e^y p_r(-y ) \ , \end{aligned}\ ] ] where @xmath70 $ ] denotes the path integral . the choice of @xmath71 and @xmath72 leading to @xmath73 preserves the involution property . hence , @xmath74 satisfies the dft , which is referred to as the crooks fluctuation theorem @xcite . the total entropy production @xmath75 corresponds to the choice of @xmath76 . in general nonequilibrium systems , the pdf @xmath52 does return to the original pdf @xmath53 under the r process . hence , the total entropy production does not obey the dft . there is an exceptional case . suppose that the external parameter @xmath20 is a time - independent constant and the system is in the nonequilibrium steady state initially . then , the system remains at the steady state at all times and the involution property holds . hence , the entropy production in the nonequilibrium steady state obeys the dft . note that the functional in ( [ y_def ] ) is the relative entropy of two probability distributions @xmath77 and @xmath78 . one may adopt the path probability distribution from a different dynamics in ( [ y_def ] ) instead of the r process . for example , if one chooses the so - called dual or adjoint dynamics , then the heat can be decomposed into two parts as @xmath79 @xcite . when a system is in a nonequilibrium steady state , it gains a work done by @xmath18 and dissipates the same amount of a heat into the heat bath . the house - keeping heat @xmath80 refers to the heat necessary to maintain the nonequilibrium steady state . the excess heat @xmath81 refers to the heat dissipated in transient dynamics @xcite . as well as the total entropy production @xmath82 , each of @xmath83 and @xmath84 is known to satisfy the ift , respectively @xcite . the total heat may be decomposed in a different way for underdamped systems @xcite , which is not covered in the paper . it is an interesting question whether the heat also satisfies the fluctuation theorem . a careful consideration of ( [ y_def ] ) provides an example with the affirmative answer . if one chooses the uniform distribution for @xmath32 and @xmath34 , the functional @xmath49 becomes equal to @xmath85 . in this case , the heat can satisfy the fluctuation theorem . however , this is a peculiar case since the uniform distribution can not be realized in systems whose phase space is unbounded . so the main question is whether the heat satisfies the fluctuation theorem in systems following the equilibrium boltzmann distribution initially . theoretical and experimental studies have shown that the heat , unlike the work , does not obey the fluctuation theorem even in the infinite @xmath86 limit @xcite . it is rather surprising because the work and the heat are proportional to @xmath2 on average while their difference @xmath87 does not scale with @xmath2 . the breakdown is attributed to a rare but large fluctuation of the system in the energy landscape @xcite . recently , a general relation , called the modified fluctuation relation , was found for the heat @xcite . it was derived for systems only with a nonconservative force . here we generalize it to cover the systems driven by both a time - dependent protocol @xmath24 and a nonconservative force . the system starts from the boltzmann distribution of ( [ p_eq ] ) with @xmath88 in the f ( r ) process so that the functional in ( [ y_def ] ) becomes as @xmath89 = \ln({{{\mathcal{p}}}[{{\bm{x}}}]}/{{{\mathcal{p}}}^r[{{\bm{x}}}^r ] } ) = \beta({{\mathcal{q}}}[{{\bm{x}}}]+\delta{{\mathcal{e}}}[{{\bm{x}}}]-\delta f)$ ] . we rewrite it as a relation between the path probabilities in the f and r processes : @xmath90 = e^{\beta({{\mathcal{q}}}[{{\bm{x}}}]+\delta{{\mathcal{e}}}[{{\bm{x}}}]-\delta f ) } \ { { \mathcal{p}}}^r[{{\bm{x}}}^r ] \ .\ ] ] multiplying both sides with @xmath91-q)\delta(\delta{{\mathcal{e}}}[{{\bm{x}}}]-e)$ ] and integrating over all paths , we obtain the fluctuation theorem @xmath92 for the joint pdf of the heat and the energy change defined as @xmath93-q ) \delta ( \delta { { \mathcal{e}}}[{{\bm{x}}}]- e ) \rangle \ .\ ] ] in deriving ( [ ft_he ] ) , we have used that @xmath94=-{{\mathcal{q}}}[{{\bm{x}}}]$ ] , @xmath95 = - \delta{{\mathcal{e}}}[{{\bm{x}}}]$ ] , and @xmath70 = \int[{d}{{\bm{x}}}^r]$ ] . the fluctuation theorem for the joint pdf was also considered in ref . @xcite . hereafter , we will use ` h ' , ` w ' , and ` e ' in the subscript for a pdf of a heat , work , and energy change , respectively . when there are multiple subscripts as in ( [ p_he ] ) , it should be understood as a joint pdf of corresponding quantities . in addition , we will use a ` @xmath96 ' in the subscript for a conditional pdf . for example , @xmath97 denotes the conditional probability that the energy change is @xmath98 given that the heat dissipation is @xmath99 . the marginal distribution of the heat is given by @xmath100 . integrating both sides of ( [ ft_he ] ) over @xmath98 , we obtain that @xmath101 the term in the parenthesis is simplified by using the conditional probability @xmath102 . it leads to the relation @xmath103 where @xmath104 applying the relation ( [ ftp_q ] ) to the r process and using the involution property that the reverse process of the r process is equivalent to the f process , one can show that @xmath105 where @xmath106 therefore , we obtain the modified fluctuation relation for the heat @xmath107 being compared with the dft in ( [ dft ] ) , the modified fluctuation relation is dressed by the additional factor @xmath108 . in order to check whether the heat satisfies the ift in ( [ ift ] ) , we evaluate @xmath109}\rangle$ ] using ( [ pfpr_he ] ) and ( [ ft_he ] ) . it yields that @xmath110-\delta f ) } \right\rangle = \left\langle e^{-\beta \delta{{\mathcal{e}}}^r[{{\bm{x}}}]}\right\rangle_r \ .\ ] ] it is also dressed by the additional factor representing the fluctuation of the energy change in the r process . the additional factor @xmath111 in ( [ dft_q ] ) reflects a correlation between the heat and the energy change during a nonequilibrium process . this calls for the study of mutual correlations between thermodynamic quantities as well as their individual fluctuations . the energy change , the work , and the heat are constrained by the thermodynamic first law in ( [ 1st_law ] ) . so , it suffices to specify a single joint pdf , e.g. , @xmath112 . the other joint pdfs are given by @xmath113 and @xmath114 from which the marginal distributions @xmath115 , @xmath116 , and @xmath117 are obtained . it is convenient to deal with the moment generating functions ( mgfs ) . we will use the symbol @xmath118 to denote a mgf , and adopt the same subscript notation for the mgf as that for the pdf . parameters @xmath119 , @xmath120 , and @xmath121 will be used as the conjugate variables for the work , the energy change , and the heat , respectively . for example , @xmath122 stands for the mgf defined as @xmath123 it is easy to show that the other mgfs for the joint pdfs are given by @xmath124 and @xmath125 the mgfs for the marginal distributions are given by @xmath126 , @xmath127 , and @xmath128 . in terms of the mgf , the dft or the crooks fluctuation theorem for the work can be written as @xmath129 or @xmath130 the modified fluctuation relation for the heat in ( [ dft_q ] ) can be rewritten as @xmath131 in comparison with ( [ ftg_w2 ] ) , the heat fluctuation deviates from the dft by the factor @xmath132 . in this section , we investigate an exactly solvable two - dimensional linear diffusion system . fluctuations and correlations are studied in detail to demonstrate the fluctuation theorems for the joint pdf and the modified fluctuation relation for the heat . especially , we focus on the question whether the dft for the heat holds asymptotically in the infinite time - interval limit or not . we investigate the nonequilibrium fluctuations in a system exerted by a linear force @xcite @xmath133 with a time - independent force matrix @xmath134 . the force is given by the sum of a conservative force @xmath135 and a nonconservative force @xmath136 with the symmetric component @xmath137 and the anti - symmetric component @xmath138 @xcite . the conservative force is also written as @xmath139 with the energy function @xmath140 . the work , heat , and energy change for a path @xmath141 are given by @xmath142 & = & - \int_{t_i}^{t_f } dt\ \dot{{{\bm{x}}}}^t(t ) \cdot { { \mathsf{f}}}_a \cdot { { \bm{x}}}(t ) \label{lds_w } \\ { { \mathcal{q}}}[{{\bm{x } } } ] & = & - \int_{t_i}^{t_f } dt\ \dot{{{\bm{x}}}}^t(t ) \cdot { { \mathsf{f}}}\cdot { { \bm{x}}}(t ) \\ \delta{{\mathcal{e}}}[{{\bm{x } } } ] & = & \frac{1}{2}{{\bm{x}}}^t_f \cdot { { \mathsf{f}}}_s \cdot { { \bm{x}}}_f - \frac{1}{2}{{\bm{x}}}^t_i \cdot { { \mathsf{f}}}_s \cdot { { \bm{x}}}_i \end{aligned}\ ] ] where the integral is of stratonovich type , @xmath143 , and @xmath144 . without time - dependent driving force , the f and r processes are the same and @xmath145 . hereafter , the energy will be measured in unit of @xmath146 and the temperature be set to unity . the system is assumed to be in the equilibrium state with the boltzmann distribution given in ( [ p_eq ] ) initially . the linear system covers a wide range of physical systems such as a colloidal particle trapped in an optical tweezer @xcite , harmonic networks @xcite , and rc circuits @xcite . in a linear diffusion system @xcite , the equilibrium relaxation dynamics and the nonequilibrium driven dynamics are competing with each other . the competition leads to an intriguing dynamical behavior . for instance , a two - dimensional linear diffusion system undergoes multiple locking - unlocking dynamical transitions with time in the tail shape of @xmath115 @xcite . in this work , we consider a simplest force matrix in two dimensions : @xmath147 with the identity matrix @xmath148 and the @xmath149 component of the pauli matrix @xmath150 . with this force matrix , the two - dimensional linear diffusion system describes a particle trapped in an isotropic harmonic potential @xmath151 and driven by a swirling nonconservative force @xmath152 of strength @xmath153 . this system does not exhibit the locking - unlocking transition observed in a system with an anisotropic harmonic potential @xcite . yet , it displays nontrivial correlations between thermodynamic quantities . in a linear diffusion system , the work fluctuation can be studied by using the path integral formalism developed in ref . we generalize the formalism to cover the joint pdf , which is explained in [ appendix ] . the formal solution for the mgf @xmath122 is given in ( [ gwe_formal ] ) . to obtain the explicit solution , one needs to solve the differential equation in ( [ dadt ] ) for @xmath154 with the initial condition @xmath155 for convenience , we will set @xmath156 hereafter . the auxiliary matrices in ( [ dadt ] ) is given by @xmath157 and @xmath158 . the off - diagonal elements of @xmath159 are anti - symmetric and @xmath160 . hence , @xmath161 , starting from the initial condition in ( [ a_ini ] ) , is proportional to the identity matrix at all @xmath21 . substituting @xmath154 with @xmath162 in ( [ dadt ] ) , one obtains a differential equation for @xmath163 : @xmath164 with @xmath165 . the solution is given by @xmath166 with @xmath167 inserting the solution into ( [ gwe_formal ] ) , one obtains @xmath168 where the function @xmath169 is defined as @xmath170 it is symmetric under @xmath171 and @xmath172 . note that it diverges when the denominator vanishes . we plot the lines of singularity at @xmath173 ( solid line ) and @xmath174 ( dashed line ) in the @xmath175 plane in fig . the solid line is obtained numerically . taking @xmath176 limit in ( [ b_def ] ) , one finds that the dashed line is parameterized as @xmath177 . it consists of the curved segments @xmath178 for @xmath179 and @xmath180 , and the linear segments @xmath181 for @xmath182 , where @xmath183 the singularity provides an information on the asymptotic tail behavior of the probability distribution @xcite , which will be analyzed further . at @xmath173 ( solid line ) and @xmath174 ( dashed line ) . ] the mgf of the work is given by @xmath184 the symmetry property of the function @xmath185 ensures the dft in ( [ ftg_w ] ) with @xmath186 . the mean value of the work is given by @xmath187 it grows linearly with time @xmath21 . that is , the driving force @xmath18 performs a work at the uniform rate @xmath188 . at a given value of @xmath21 , @xmath189 has simple poles at @xmath190 and @xmath191 , which are marked with the closed circles in fig . [ fig1 ] . due to the symmetry @xmath192 , @xmath193 . the simple poles indicate exponential tails in @xmath115 : @xmath194 e^{\lambda_-(t)w } & , & w \gg 1 \ . \end{array}\right.\ ] ] the tail shape of @xmath115 can be characterized more quantitatively by the large deviation functions ( ldfs ) @xcite defined as @xmath195 the symbol @xmath196 is used to denote the ldf for the pdf ( mgf ) . they are related through the legendre transformation @xmath197 by taking the @xmath176 limit of ( [ gw_solution ] ) , we obtain that @xmath198 -\infty & , & \mbox{otherwise}\end{array}\right . \ .\ ] ] so the legendre transformation yields that @xmath199 it has the limiting behavior @xmath200 , which is consistent with the exponential tails in ( [ tail_pw ] ) . in terms of the ldf , the dft for the work is written as @xmath201 and @xmath202 . the explicit solutions in ( [ ldf_gw ] ) and ( [ ldf_pw ] ) confirm the dft . the mgf of the energy change is given by @xmath203 it also has simple poles at @xmath204 with @xmath205 , which are marked with the closed diamonds in fig . the simple poles indicate that @xmath206 for large @xmath207 . one can obtain @xmath117 exactly from the fourier transform @xmath208 it indeed decays exponentially . the ldf is given by @xmath209 the energy change is distributed symmetrically . so , the system may gain or lose the energy equally probably . it is interesting to note that the pdf is independent of the driving strength @xmath153 . the independence is due to a specific property of the force matrix in ( [ f_solvable ] ) . the nonconservative force @xmath210 is perpendicular to the conservative force @xmath211 . hence the nonequilibrium force does not affect the energy fluctuation . the mgf of the heat is given by @xmath212 it has simple poles at @xmath213 , which are marked with closed squares in fig . [ fig1 ] . from the @xmath176 limit of ( [ geta ] ) , the ldf is given by @xmath214 -\infty & , & \mbox{otherwise } \end{array}\right.\ ] ] where @xmath215 and @xmath216 - \frac{1}{2}\left(\sqrt{1 + 1/\varepsilon^2}-1\right ) & , & \varepsilon^2>1/3 \end{array}\right.\ ] ] note that @xmath217 has the same function form as @xmath218 , but @xmath219 is supported in the narrower domain . we compare the two ldfs in fig . [ fig2](a ) and ( b ) . it is straightforward to obtain the ldf @xmath220 from the legendre transformation @xmath221 : @xmath222 \sqrt{\frac{(1+\varepsilon^2)(q^2 + 4\varepsilon^2 ) } { 4\varepsilon^2}}-\frac{q}{2}-1 & , & q_+^ * \leq q \leq q_-^ * \\ [ 2 mm ] \frac{1-\varepsilon^2}{1+\varepsilon^2 } q - \frac{4\varepsilon^2}{1+\varepsilon^2 } & , & q_-^ * < q \end{array}\right.\ ] ] where @xmath223 and @xmath224 the ldf for the heat is compared with that of the work in fig . [ fig2](c ) and ( d ) . they deviate from each other at large values of @xmath225 and @xmath226 . the heat exhibits stronger fluctuations than the work . . the parameter values are @xmath227 in ( a ) , ( c ) , and ( e ) , and @xmath228 in ( b ) , ( d ) , and ( f ) . ] we test whether the ldf of the heat satisfies the dft . the modified fluctuation relation in ( [ dft_q ] ) is rewritten as @xmath229 where @xmath230 with @xmath111 in ( [ psi_def ] ) . using ( [ ldf_q ] ) , we find that @xmath231 \frac{q}{2 } - \frac{\sqrt{(1+\varepsilon^2)(q^2 + 4\varepsilon^2)}}{2\varepsilon}+1 & , & -q_+^ * \leq q < q_-^ * \\ [ 2 mm ] -\frac{1-\varepsilon^2}{1+\varepsilon^2}q + \frac{4\varepsilon^2}{1+\varepsilon^2 } & , & q_-^\leq q \end{array}\right.\ ] ] and @xmath232 . it is plotted in fig . [ fig2](e ) and ( f ) . the heat appears to obey the dft with @xmath233 within the interval @xmath234 . however , rare fluctuations with large values of @xmath226 do not obey the fluctuation theorem . we have shown that the heat does not obey the dft even in the @xmath176 limit . the correction factor @xmath235 calculated in ( [ psi_res ] ) does not vanish in the large @xmath226 region . note that @xmath236 reflects the mutual correlation between the heat and the energy change in nonequilibrium dynamics . in this subsection , we investigate the mutual correlations among them . the ldf for the joint distribution of @xmath0 and @xmath4 is defined by @xmath237 using the expression in ( [ gwe_lin_sol ] ) and the analytic property of the function @xmath185 defined in ( [ b_def ] ) , we obtain that @xmath238 -\infty & , & \mbox{otherwise}\end{array}\right.\ ] ] where @xmath239 denotes the domain bounded by the dashed line in fig . the legendre transformation @xmath240 yields that @xmath241 it is minimum at @xmath242 and increases linearly in @xmath226 and @xmath243 asymptotically . the joint distribution of @xmath1 and @xmath4 is related to that of @xmath0 and @xmath4 through ( [ he_we ] ) and ( [ g_he ] ) . hence , the ldfs @xmath244 for @xmath245 and @xmath246 for @xmath247 are given by @xmath248 in ( a ) and @xmath246 in ( b ) at @xmath249 . also drawn are the equiprobable lines . ] we compare @xmath250 and @xmath246 at @xmath249 with the density plot in fig . [ fig3 ] . interestingly , @xmath251 at all values of @xmath252 . irrespective of an external work , the internal energy of the system may increase or decrease with the equal probability . by contrast , a strong anti - correlation exists between the heat dissipation and the energy change . the system tends to lose an internal energy when it dissipates a heat and to gain an internal energy when it absorbs a heat . the mutual correlation is represented well by the most probable value of the energy change @xmath253 to a given value of @xmath254 . that is to say , @xmath246 is minimum at @xmath255 when it is regarded as a function of @xmath256 with fixed @xmath254 . using the explicit solution for @xmath246 in ( [ pi_we ] ) and ( [ pi_he ] ) , we find that @xmath257 0 & , & q_+^ \leq q < q_-^ * \\ [ 2 mm ] \frac{-(1 - 3\varepsilon^2)q+2\varepsilon^2(3-\varepsilon^2)}{1-\varepsilon^4 } & , & q > q_-^ * \end{array}\right.\ ] ] it takes a positive value for @xmath258 and a negative value for @xmath259 . we will show that the nonzero @xmath253 is directly related to the breakdown of the dft for the heat . the dft can be examined by checking whether the relation @xmath260 holds or not . note that @xmath261\sim \exp[-t \min_{e } \{\pi_{\rm he}(q / t , e)\}]$ ] , which yields that @xmath262 . then , it follows that @xmath263 \\ & & - \left[\pi_{\rm he}(-q , e_{pr}(-q))-\pi_{\rm he}(-q,0)\right ] \ .\end{aligned}\ ] ] comparing this relation with ( [ dft_ldf_q ] ) and using @xmath264 , we obtain an alternative expression for @xmath235 : @xmath265 \nonumber \\ & - & \left[\pi_{\rm he}(-q , e_{pr}(-q))-\pi_{\rm he}(-q,0)\right ] \ .\end{aligned}\ ] ] it would vanish if @xmath253 were equal to 0 at all @xmath254 . the negative correlation between the heat and the energy change leads to nonzero values of @xmath253 , hence nonzero values of @xmath235 . this analysis shows that the correlation is the origin of the breakdown of the dft for the heat . we have derived the modified fluctuation relation for the heat given in ( [ dft_q ] ) in general overdamped langevin systems . it involves the extra factor @xmath111 defined in ( [ psi_def ] ) which reflects the correlation between the dissipated heat and the energy change of the system . we have investigated the mutual correlations in a two - dimensional linear diffusion system , which is exactly solvable by using the method presented in [ appendix ] . we have obtained the closed form expressions for the ldfs for the probability distributions in ( [ ldf_pw ] ) , ( [ ldf_e ] ) , ( [ ldf_q ] ) , ( [ pi_we ] ) , and ( [ pi_he ] ) . in particular , the result shows that the heat and the energy change are negatively correlated . it is manifested in @xmath253 that corresponds to the value of @xmath256 minimizing @xmath246 with fixed @xmath254 , i.e. , the most probable energy change to a given heat dissipation . it is nonzero for large values of @xmath254 as shown in ( [ e_pr_res ] ) , and gives rise to a nonzero value of @xmath235 . therefore , we conclude that the negative correlation is the origin for the distinct fluctuation property of the heat . in this work , we only study the simplest linear diffusion system with the force matrix given in ( [ f_solvable ] ) . a natural direction for future works is to consider a general force matrix having different diagonal elements , which can describe experimental systems such as a coupled rc circuit investigated in ref . . it will be interesting to study the correlations in the anisotropic system . it will be also interesting to study the correlations in nonlinear systems . we hope that our work triggers further theoretical and experimental studies . we would like to thank hyunggyu park , chulan park , and jong - min park for helpful discussions . this work was supported by the basic science research program through the nrf grant no . in this appendix , we introduce a path integral formalism for the mgf @xmath266 in a linear diffusion system where the force is given by the form in ( [ f_lin ] ) . it is an extension of the path integral formalism for @xmath267 developed in ref . @xcite . suppose that the initial configuration @xmath268 follows the distribution given by @xmath269 with a symmetric positive - definite matrix @xmath270 . the equilibrium boltzmann distribution is obtained by taking @xmath271 . in general , it can be taken arbitrarily depending on a physical condition . the mgf @xmath189 is given by @xcite @xmath272 { { \mathcal{t}}}[{{\bm{x}}}(t)\|{{\bm{x}}}(t_i ) ] p_i({{\bm{x}}}(t_i ) ) e^{-\lambda { { \mathcal{w}}}[{{\bm{x}}}(t ) ] } \ , \ ] ] where @xmath273 $ ] is the conditional probability for a path @xmath25 to a given initial configuration @xmath268 and the work is given in ( [ lds_w ] ) . the onsager - machlup formalism @xcite allows one to write the conditional probability as @xmath274 \propto e^{-\frac{1}{4}\int_{t_i}^{t_f } dt \| \dot{{\bm{x}}}(t)-{{\bm{f}}}({{\bm{x}}}(t))\|^2 - \frac{1}{2}\bm\nabla_{{{\bm{x } } } } \cdot { { \bm{f}}}(t)}\ ] ] with the stratonovich integral . note that the exponents in ( [ lds_w ] ) , ( [ p0 ] ) , and ( [ path_prob ] ) are quadratic in @xmath55 . hence , @xmath189 in ( [ gw ] ) can be evaluated by the gaussian integration in principle . the gaussian integration can be evaluated efficiently . first , discretize the time and introduce @xmath275 with @xmath276 representing the configuration at @xmath277-th time slice . then , the mft in ( [ gw ] ) is written in the form of @xmath278 e^{-{{\bm{x}}}_0^t \cdot { { \mathsf{s}}}\cdot { { \bm{x}}}_0 } \ , \ ] ] where @xmath279 is quadratic in both @xmath280 and @xmath281 that is determined from the expressions for @xmath282 and @xmath0 @xcite . the kernel for @xmath283 is @xmath270 . since @xmath283 is coupled to @xmath284 through @xmath285 , one obtains an effective kernel for @xmath284 after integrating over @xmath283 . it is done successively to obtain a recursion relation for the kernel , denoted by @xmath286 , for @xmath287 . in the @xmath288 limit , the recursion relation can be casted into the differential equation @xmath289 where @xmath290 and @xmath291 . the initial condition is given by @xmath292 collecting all the factors coming from the all integrations , one obtains that @xmath293 the factor @xmath294 originates from the normalization factor of @xmath32 in ( [ p0 ] ) and the factor @xmath295 from the final gaussian integration over @xmath296 . the exponential factor accounts for the contribution from the intermediate time steps @xcite . the formal expression for @xmath297 is given by @xmath298 { { \mathcal{t}}}[{{\bm{x}}}(t)\|{{\bm{x}}}(t_i ) ] p_i({{\bm{x}}}(t_i ) ) \nonumber \\ & \times & e^{-\lambda{{\mathcal{w}}}[{{\bm{x}}}(t)]-\kappa \delta{{\mathcal{e}}}[{{\bm{x}}}(t))]}\end{aligned}\ ] ] note that @xmath299 $ ] is also quadratic in @xmath55 . hence , @xmath300 can be evaluated using the same method . comparing the integrands in ( [ gw ] ) and ( [ gwe ] ) , one finds that they differ by the boundary term @xmath301 = -\frac{\kappa}{2 } { { \bm{x}}}(t_f)^t \cdot { { \mathsf{f}}}_s \cdot { { \bm{x}}}(t_f ) + \frac{\kappa}{2 } { { \bm{x}}}(t_i)^t \cdot { { \mathsf{f}}}_s \cdot { { \bm{x}}}(t_i)$ ] in the exponent . consequently , @xmath302 is simply obtained by replacing @xmath270 in ( [ dadt_initial ] ) with @xmath303 and @xmath304 in ( [ gw_formal ] ) with @xmath305 . therefore , we obtain @xmath306 where @xmath154 is the solution of ( [ dadt ] ) with the shifted initial condition @xmath307 the initial state of the system at time @xmath54 is characterized by @xmath270 . if one takes the equilibrium boltzmann distribution as the initial state , it should be taken as @xmath308 99 d.j . evans , e.g.d . cohen , and g.p . morriss , phys . rev . lett . * 71 * , 2401 ( 1993 ) . g. gallavotti and e.g.d . cohen , phys . lett . * 74 * , 2694 ( 1995 ) . c. jarzynski , phys . lett . * 78 * , 2690 ( 1997 ) . crooks , phys . e * 60 * , 2721 ( 1999 ) . j. kurchan , j. phys . a * 31 * , 3719 ( 1998 ) . j.l . lebowitz and h. spohn , j. stat . phys . * 95 * , 333 ( 1999 ) . t. hatano and s .- sasa , phys . lett . * 86 * , 3463 ( 2001 ) . u. seifert , phys . * 95 * , 040602 ( 2005 ) . m. esposito and c. van den broeck , phys . lett . * 104 * , 090601 ( 2010 ) . t. sagawa and m. ueda , phys . * 104 * , 090602 ( 2010 ) . spinney and i.j . ford , phys . lett . * 108 * , 170603 ( 2012 ) . lee , c. kwon , and h. park , phys . 110 * , 050602 ( 2013 ) . wang , e.m . sevick , e. mittag , d.j . searles , and d. j. evans , phys . 89 * , 050601 ( 2002 ) . g. hummer , a. szabo , proc . usa * 98 * , 3658 ( 2001 ) . j. liphardt , s. dumont , s.b . smith , i. tinoco jr . , and c. bustamante , science * 296 * , 1832 ( 2002 ) . k. hayashi , h. ueno , r. iino , and h. noji , phys . * 104 * , 218103 ( 2010 ) . s. ciliberto , s. joubaud , and a. petrosyan , j. stat . ( 2010 ) p12003 . u. seifert , rep . . phys . * 75 * , 126001 ( 2012 ) . r. garca - garca , d. domnguez , v. lecomte , and a.b . kolton , phys . e * 82 * , 030104(r ) ( 2010 ) . r. garca - garca , v. lecomte , a. b. kolton , and d. domnguez , j. stat . ( 2012 ) p02009 . noh and j .- m . park , phys . lett . * 108 * , 240603 ( 2012 ) . j. farago , j. stat . phys . * 107 * , 781 ( 2002 ) . r. van zon and e.g.d . cohen , phys . 91 * , 110601 ( 2003 ) . r. van zon and e.g.d . cohen , phys . rev e * 69 * , 056121 ( 2004 ) . n. garnier and s. ciliberto , phys . e * 71 * , 060101(r ) ( 2005 ) . p. visco , j. stat . ( 2006 ) p06006 . harris , a. rkos , and g.m . schtz , europhys . lett . * 75 * , 227 ( 2006 ) . a. rkos and r.j . harris , j. stat . ( 2008 ) p05005 . fogedby and a. imparato , j. stat . ( 2011 ) p05015 . t. nemoto , phys . rev . e * 85 * , 061124 ( 2012 ) . a. puglisi , l. rondoni , and a. vulpiani , j. stat . ( 2006 ) p08010 . lee , c. kwon , and h. park , phys . e * 87 * , 020104(r ) ( 2013 ) . c. kwon , j.d . noh , and h. park , phys . e * 83 * , 061145 ( 2011 ) . noh , c. kwon , and h. park , phys . lett . * 111 * , 130601 ( 2013 ) . c. kwon , j.d . noh , and h. park , phys . e * 88 * , 062102 ( 2013 ) . gardiner , _ handbook of stochastic methods for physics , chemistry and the natural sciences _ , 2nd ed . ( springer , berlin , 1985 ) . h. risken , _ the fokker - planck equation _ ( springer , berlin , 1989 ) . k. sekimoto , prog . 130 * , 17 ( 1998 ) . t. m. cover and j. a. thomas , _ elements of information theory _ ( wiley , new york , 1991 ) . l. onsager and s. machlup , phys . rev . * 91 * , 1505 ( 1953 ) . in general , one may decompose the force matrix as @xmath309 with an _ arbitrary symmetric _ matrix @xmath310 . with any choice of @xmath310 , one can define the energy and the nonequlibrium work consistently as was done in ref . @xcite . in this work , we only consider the specific but natural choice of @xmath311 in order to avoid an unnecessary complication . a. kundu , s. sabhapandit , and a. dhar , j. stat . ( 2011 ) p03007 . k. saito and a. dhar , phys . e * 83 * , 041121 ( 2011 ) . s. sabhapandit , phys . e * 85 * , 021108 ( 2012 ) . fogedby and a. imparato , j. stat . ( 2012 ) p04005 . s. ciliberto , a. imparato , a. naert , and m. tanase , phys . rev . lett . * 110 * , 180601 ( 2013 ) . h. touchette , phys . rep . * 478 * , 1 ( 2009 ) . a.w.c . lau and t.c . lubensky , phys . e * 76 * , 011123 ( 2007 ) .	nonequilibrium systems exchange the energy with an environment in the form of work and heat . the work done on a system obeys the fluctuation theorem , while the dissipated heat which differs from the work by the internal energy change does not . we derive the modified fluctuation relation for the heat in the overdamped langevin system . it shows that mutual correlations among the work , the heat , and the internal energy change are responsible for the different fluctuation property of the work and the heat . the mutual correlation is investigated in detail in a two - dimensional linear diffusion system . we develop an analytic method which allows one to calculate the large deviation function for the joint probability distributions . we find that the heat and the internal energy change have a negative correlation , which explains the reason for the breakdown of the fluctuation theorem for the heat .
( 2380 ) dibaryon resonance signatures in recent wasa - at - cosy collaboration experiments . left : from the peak observed in the @xmath10 reaction @xcite . right : from the argand diagram of the @xmath11 partial wave in @xmath12 scattering @xcite.,title="fig:",scaledwidth=48.0%,height=188 ] ( 2380 ) dibaryon resonance signatures in recent wasa - at - cosy collaboration experiments . left : from the peak observed in the @xmath10 reaction @xcite . right : from the argand diagram of the @xmath11 partial wave in @xmath12 scattering @xcite.,title="fig:",scaledwidth=48.0%,height=188 ] the wasa - at - cosy collaboration observed a peak about 80 mev below the @xmath3 threshold in the @xmath10 reaction @xcite . this peak , shown on the left panel of fig . [ fig : wasa ] , was identified with the @xmath13=@xmath14 @xmath15 @xmath3 dibaryon predicted in 1964 by dyson and xuong @xcite . somewhat surprisingly , the width @xmath16 mev is considerably smaller than even the width of a single @xmath17 baryon , @xmath18 mev . the @xmath19 isospin assignment follows from the isospin balance in @xmath20 and the @xmath21 spin - parity assignment follows from the measured decay - deuteron angular distribution . the @xmath2(2380 ) was also observed in @xmath22 , with cross section consistent with the one measured in @xmath10 @xcite . several other related @xmath23 reactions were also studied @xcite . recent measurements of @xmath12 scattering and analyzing power @xcite have led to the @xmath12 @xmath11 partial - wave argand diagram shown on the right panel of fig . [ fig : wasa ] , supporting the @xmath2(2380 ) dibaryon resonance interpretation @xcite . the @xmath2(2380 ) width and decay branching ratios have been discussed recently in several quark - based calculations @xcite . the term ` quark - based ' does not necessarily mean that the resulting @xmath2(2380 ) is of a purely _ hexaquark _ structure . in fact , a recent quark - model study of spatially symmetric @xmath24 @xmath25 states finds the @xmath13=@xmath14 hexaquark several hundreds of mev above the @xmath3 threshold @xcite . it is by adding potentially double - counting meson exchanges , e.g. a scalar - isoscalar @xmath26 meson , and applying resonating group methods ( rgm ) , that quark - based calculations generate a tightly bound @xmath3 dibaryon . in this note we argue that the observed @xmath2(2380 ) resonance energy and width are incompatible with a leading @xmath3 asymptotic structure . using the gal - garcilazo hadronic model that views @xmath2(2380 ) as a @xmath7 resonance embedded in the @xmath4 continuum some 80 mev below the @xmath3 threshold @xcite , we discuss its width and decay branching ratios in comparison to those deduced from experiment by bashkanov et al . in particular , predictions are made for the @xmath8 partial decay widths which are suppressed to leading order within a purely @xmath3 description of @xmath2(2380 ) . we also comment on a possible connection of the abc effect observed in @xmath27 @xcite to the @xmath2(2380 ) dibaryon . we begin with a brief overview of these hadronic - model calculations , followed by discussion of the @xmath2(2380 ) width and decay branching ratios . @xmath28 and @xmath3 @xmath29-wave dibaryon resonances @xmath30 with isospin @xmath31 and spin @xmath32 were proposed by dyson and xuong @xcite who focused on the lowest - dimension su(6 ) multiplet in the @xmath33 product that contains the su(3 ) @xmath34 and @xmath35 multiplets in which the deuteron @xmath36 and @xmath6 virtual state @xmath37 are classified . this yields two non - strange dibaryon candidates , @xmath38 ( @xmath28 ) and @xmath39 ( @xmath3 ) . identifying the constant @xmath40 in the resulting spin - isospin su(4 ) mass formula @xmath41 \label{eq : dyson}\ ] ] with the @xmath6 threshold mass 1878 mev , a value @xmath42 mev was derived by assigning @xmath38 to the @xmath43 resonance at @xmath44 mev , near the @xmath28 threshold , which had been observed during the 1950s @xcite , thereby leading to the prediction @xmath45=2350 mev . -matrix pole equation for the @xmath39 @xmath3 dibaryon @xcite.,scaledwidth=60.0% ] in the absence of four - body @xmath46 calculations of @xmath3 dibaryons , the @xmath39 dibaryon was studied in ref . @xcite , instead , within a @xmath7 three - body model in which the @xmath47 interaction is limited to the dominant @xmath17-resonance @xmath48 channel and the @xmath28 interaction is limited to the @xmath49 @xmath38 dibaryon channel . the faddeev equations of this @xmath7 three - body model give rise to an effective @xmath3 lippmann - schwinger ( ls ) equation shown diagrammatically in fig . [ fig : d03 ] , where @xmath50 stands for the @xmath38 dibaryon and @xmath51 denotes a stable @xmath17(1232 ) . these faddeev equations lead equally well to an effective @xmath52 ls equation obtained by substituting @xmath53 , @xmath54 and @xmath55 in the @xmath3 ls equation of fig . [ fig : d03 ] . thus , in this @xmath7 three - body model , the @xmath2(2380 ) dibaryon is a @xmath52@xmath3 coupled - channels resonance . for a more complete discussion of @xmath30 @xmath3 dibaryon candidates , see ref . @xcite . .@xmath32-matrix poles @xmath56 and @xmath57 of the lowest @xmath28 and @xmath3 dibaryons obtained in @xmath58 and @xmath7 faddeev calculations @xcite using a @xmath47 @xmath48 form factor of size 1.35 fm , a spectator-@xmath51 complex mass @xmath59 in the propagator of the ls equation depicted in fig . [ fig : d03 ] , and @xmath60 interaction that resonates in the @xmath49 channel at @xmath61 , all in mev . [ cols="^,^,^,^",options="header " , ] [ tab : br ] comparing the br obtained in the three model versions specified by their value of the @xmath3 fraction @xmath62 with those derived from experiment in ref . @xcite and listed in the last column of table [ tab : br ] , one notes the similarity between the br obtained in a purely @xmath3 model ( @xmath63 ) and those derived from experiment . in fact , this similarity is somewhat fortuitous because all three model versions were designed to reproduce the @xmath2 peak cross sections derived from experiment , @xmath64 @xmath65b and @xmath66 @xmath65b @xcite , thereby agreeing also for the rest of the @xmath67 cross sections . the three model versions figuring in table [ tab : br ] differ essentially _ only _ in their @xmath5 br which is obviously zero in the purely @xmath3 model ( @xmath63 ) . the @xmath5 partial decay width and br listed for the purely @xmath52 model ( @xmath68 ) were normalized to a total @xmath2 pionic width of 75@xmath6911=64 mev . the relatively high value of @xmath7015% for the obtained br is excluded by a recent determination of a @xmath71 upper limit @xcite . in contrast , a value of the @xmath5 br smaller by about a factor of two was obtained , by applying the @xmath52 decay fraction @xmath72 to the @xmath6 partial decay width @xmath73 mev , in the specific mixing model version listing in table [ tab : br ] . how robust are the br results shown for the @xmath3@xmath52 coupled channels scheme in table [ tab : br ] ? the listed br are based on assuming a value @xmath74 mev . a @xmath7510% variation of this value results in @xmath7650% variation in @xmath77 away from its initially assumed value which can be restored by merely a @xmath7510% variation in @xmath78 away from its initially assumed value . we conclude that the partial decay widths and br listed in table [ tab : br ] for a mixing parameter @xmath79 have uncertainties of up to 10% , except for those for the @xmath5 decay mode which depend only on the assumed value of @xmath62 . next we allow @xmath62 to vary by assuming a @xmath7510 mev uncertainty in the value of @xmath80 mev that served as input through eq . ( [ eq : alpha ] ) for deriving the value of @xmath79 for use in table [ tab : br ] . the resulting values for @xmath81 and 50 mev are @xmath82 and @xmath83 , respectively , leading to the following uncertainty estimate for the @xmath5 decay mode : @xmath84 the @xmath2(2380 ) is the most promising dibaryon candidate at present , supported by systematic studies of its production and decay in recent wasa - at - cosy experiments @xcite . in most theoretical works , beginning with the 1964 dyson - xuong prediction @xcite , it is assigned as a @xmath3 quasibound state . given the small width @xmath85 mev with respect to twice the width of a free - space @xmath17 , @xmath18 mev , its location far from thresholds makes it easier to discard a possible underlying threshold effect . however , as argued in this work , the observed small width is much larger than what two _ deeply bound _ @xmath17 baryons can yield upon decay . the @xmath86(2380 ) therefore can not be described exclusively by a @xmath3 component . a complementary quasi two - body component is offered in the @xmath87 three - body hadronic model of refs . @xcite by a @xmath52 channel , in which the @xmath2(2380 ) resonates . the @xmath38 dibaryon stands here for the @xmath49 @xmath28 near - threshold system that might or might not possess a quasibound state @xmath32-matrix pole . it is a loose system of size typically 1.52 fm , as opposed to the compact @xmath3 component of size 0.51 fm . it was pointed out how the abc low - mass enhancement in the @xmath88 invariant mass distribution of the @xmath10 fusion reaction at @xmath89 gev might be associated with the small size of the @xmath3 component . furthermore , we have shown how to determine the relative weight of these two components by fitting to the total @xmath2(2380 ) width , thereby deriving @xmath2 partial decay widths and branching ratios that agree with experiment @xcite . a new element in the present derivation is the ability , through the @xmath52 channel , to evaluate the @xmath90 decay width and br . our prediction is for a br of order 7% , consistent with an upper limit of @xmath71 determined recently @xcite . special thanks are due to heinz clement for stimulating exchanges on the physics of dibaryons @xcite , and to maria platonova on the @xmath52 interpretation of the @xmath2(2380 ) @xcite . here we demonstrate that the isospin structure @xmath91_{i=0}$ ] of the @xmath2(2380 ) dibaryon resonance is the same for its @xmath3 component and for its @xmath92 component provided the angular momentum couplings involved are unique , which holds for the leading configurations @xmath93 and @xmath94 of @xmath38 and @xmath2 , respectively . starting with a @xmath95 structure of @xmath2(2380 ) , we recall that @xmath38 stands for a near - threshold @xmath96 dibaryon , where @xmath17 is a @xmath48 @xmath97 resonance , and recouple in isospace : @xmath98_{i_{{\cal d}_{12 } } } \otimes\pi_2\right\}_{i=0}\to \left\{[(n_1\otimes n_2)_{i_{nn}}\otimes \pi_1 ] _ { i_{{\cal d}_{12 } } } \otimes \pi_2\right\}_{i=0 } , \label{eq : iso1}\ ] ] using the @xmath99 orthogonal transformation with elements given by @xmath100 where @xmath101 and @xmath102 denote the nucleons and pions isospins , respectively , and @xmath103 . the square of the @xmath104 element is 2/3 , and that of the @xmath105 element is 1/3 . the @xmath104 projection factor 2/3 is the same as that considered within the @xmath3 model in section [ sec : s=0 ] . note that the @xmath5 faddeev calculation of @xmath38 in refs . @xcite was based on a single @xmath104 , @xmath106 @xmath29-wave configuration , thereby justifying the @xmath104 projection applied here . note also that with @xmath106 and two @xmath107-wave pions , the angular momentum coupling needed for the @xmath108 @xmath2(2380 ) is unique , with all individual components parallel to each other . therefore we need to focus just on the isospin recoupling . proceeding to recouple the state on the r.h.s . of eq . ( [ eq : iso1 ] ) , @xmath109_{i_{{\cal d}_{12}}}\otimes \pi_2\right\}_{i=0}\to[(n_1\otimes n_2)_{i_{nn}}\otimes ( \pi_1 \otimes \pi_2)_{i_{\pi\pi}}]_{i=0 } , \label{eq : iso3}\ ] ] where @xmath110 , we get @xmath99 transformation elements identically 1 , @xmath111 irrespective of the value of @xmath112 . in refs . @xcite we got the state on the r.h.s . ( [ eq : iso3 ] ) by recoupling directly from a @xmath3 configuration , @xmath113_{i=0 } \to [ ( n_1\otimes n_2)_{i_{nn}}\otimes ( \pi_1 \otimes \pi_2)_{i_{\pi\pi}}]_{i=0 } , \label{eq : iso5}\ ] ] using the @xmath114 transformation @xmath115 which yields precisely the same @xmath99 transformation elements as in eq . ( [ eq : iso2 ] ) . this establishes the equivalence of the @xmath92 and @xmath3 bases as far as the leading contributions to the @xmath116 branching ratios calculated in section [ sec : br ] are concerned . p. adlarson et al . ( wasa - at - cosy collaboration ) , phys . rev . lett . * 106 * ( 2011 ) 242302 . see also the preceding reports : h. clement et al . ( celsius - wasa collaboration ) , prog . * 61 * ( 2008 ) 276 ; m. bashkanov et al . ( celsius / wasa collaboration ) , phys . * 102 * ( 2009 ) 052301 . p. adlarson et al . ( wasa - at - cosy collaboration , said data analysis center ) , phys . c * 90 * ( 2014 ) 035204 . see also p. adlarson et al . ( wasa - at - cosy collaboration , said data analysis center ) , phys . * 112 * ( 2014 ) 202301 .	attempts to reproduce theoretically the relatively small width , @xmath0 mev , of the @xmath1 @xmath2(2380 ) dibaryon resonance established by the wasa - at - cosy collaboration are discussed . the validity of associating the @xmath2(2380 ) in quark - based models primarily with a tightly bound @xmath3 configuration is questioned . the @xmath2(2380 ) width and decay branching ratios into @xmath4 , @xmath5 and @xmath6 final states are studied within the gal - garcilazo hadronic model in which the @xmath2(2380 ) is a @xmath7 resonance embedded in the @xmath4 continuum some 80 mev below the @xmath3 threshold . in particular , predictions are made for the branching ratios of the unobserved yet @xmath8 decays which are suppressed in a purely-@xmath3 dibaryon model . comments are also made on a possible connection of the abc effect observed in the @xmath9 resonance reaction to the @xmath2(2380 ) dibaryon . pion - assisted dibaryons ; @xmath2(2380 ) @xmath3 dibaryon
conditional measurements in the atom field interaction @xcite , i.e. projections of the entangled state onto a selected atomic state once the atom and field have interacted and the atom has exit the cavity , carrying the information of the field inside it , have been used to produce nonclassical states of light such as schrdinger cat states @xcite or more complex superpositions of coherent states @xcite that may lead to generation of numbers states . gosh and gerry @xcite have shown that a class of nonclassical states may be produced by such conditional measurements which luis @xcite called paradoxical , as he showed that a measurement leading to absorption of light produced an increase of the number of photons . more recently , valverde and baseia @xcite have studied such paradoxical generation of nonclassical states in models of interpolating hamiltonians , in particular of the ( nonlinear ) jaynes - cummings type . here we show that conditional measurements can lead efficiently to multiphoton processes @xcite , and in this way , a resonant interaction that allows single - photon exchange of energy between atom and field @xcite may generate , via highly probable conditional measurements , the creation or annihilation of more than one photons in the cavity . because the atom - field interaction may be modeled in other type of systems , namely ion - laser interactions @xcite , the results shown here may be translated to such systems , where one could use this effect to cool trapped ions . consider the interaction hamiltonian for the on - resonance interaction of an atom and a quantized field ( @xmath0 is set to one ) @xmath1 the operators @xmath2 and @xmath3 are the annihilation and creation operators , the parameter @xmath4 is the so - called rabi frequency and the atomic operators , @xmath5 and @xmath6 , are the lowering and raising pauli matrices , respectively . we can obtain the evolution operator , @xmath7 , by expanding the exponential in taylor series ; the even and odd powers may be written as ( we use the pauli matrices in their @xmath8 matrix notation ) @xcite @xmath9 and @xmath10 with @xmath11 the london phase operator @xcite . so , we may obtain the evolution operator in the form @xmath12 if we consider an initial coherent state for the field , @xmath13 , with its distribution in terms of number states given by @xmath14 are the fock states , and an excited state for the atom , the initial wavefunction is given by @xmath15 we obtain the evolved wavefunction as @xmath16 , that in the ( complete ) ket notation is written as @xmath17 and @xmath18 are given by @xmath19 may be written as a schmidt decomposition @xcite @xmath20 are the eigenvalues of the atomic density matrix an the orthonormal field wavefunctions are written as @xmath21 the quantities @xmath22 , @xmath23 , @xmath24 and @xmath25 are related as follows @xmath26 and @xmath27 by measuring the atom as it exits the cavity in the @xmath28 state , i.e. doing a conditional measurement of the atom in that specific schmidt state , the field is projected to the normalized state @xmath29 . from it , we can calculate the average number of photons , @xmath30 . we plot this quantity as a function of @xmath31 in figure 1 , where it is shown that an increase ( after a severe decrease ) of the average number of photons would occur for certain times ( where the maximums happens ) . of course , although it may be seen as a surprising effect such nonlinear generation of photons in the cavity , it may be understood by the fact that a distribution of photons is initially considered as well as the fact that the conditional measurement is in itself a non - unitary process . . the field is initially in a coherent state with ( a ) @xmath32 , ( b ) @xmath33 and ( c ) @xmath34 . ] in figure 2 , we plot ( dashed line ) the probability of such events occurring . we see that for the first maximum of the solid line in figure 1 , @xmath35 , the probability shown in figure 2 is a minimum yielding a probability of approximately @xmath36 percent . it may be also seen that all the maximums happen with minimum probability , although this probability increases such that the second relevant multiphoton processhas a probability to occur greater than 20 percent . ) of measuring the atom after it exits the cavity in the ground state ( solid line ) or excited state ( dash line ) . ] in figure 3 , we plot the mandel-@xmath37 parameter @xcite @xmath38 as a function of @xmath31 where the generation of nonclassical states may be seen for times for which the mandel-@xmath37 parameter is less than zero . parameter as a function of @xmath31 for the state @xmath29 the field is initially in a coherent state with ( a ) @xmath32 , ( b ) @xmath33 and ( c ) @xmath34 . ] in figure 4 ( b ) , we plot the photon distribution , @xmath39 as a function of @xmath40 for @xmath32 and @xmath35 . a deviation from the original distribution , figure 4 ( a ) , centered at @xmath41 may be seen , as the one obtained from a conditional measurement may be seen centered at about @xmath42 . therefore , an increase of the average number of photons greater than one , i.e. a multiphoton process , takes place . this figure resembles that of a squeezed state @xcite . in fact , squeezed light has been predicted in the jaynes - cummings model by kuklinski and madajczyk @xcite and its interaction with two - level atoms studied by satyanarayana _ _ @xcite and moya - cessa and vidiella - barranco @xcite . and ( b ) the conditional state @xmath29 , with @xmath32 and an interaction time @xmath43 . ] or @xmath44 state , respectively . ] we have shown that , after an atom resonantly interacts with a field , the average number of photons inside the cavity is severely affected by measuring it when it exits it . we call this a multiphoton process as the atom - field interaction allows the exchange of only single photons , but the conditional measurement generates the creation or absorption of several photons in the cavity . we have shown , numerically , that such effects are related to the initial number of photons in the cavity for both coherent states and thermal distributions giving rise to the addition or subtraction proportional to the square of the initial average number of photons . 999 sherman b. ; kurizki g. _ phys . rev . _ * 1992 * , a 45 , r7674 . vogel k. ; akulin v. m. ; schleich w.p . lett . _ * 1993 * , 71 , 1816 . garraway b.m . ; sherman b. ; moya - cessa h. ; knight p.l . ; kurizki g. _ phys . rev . _ * 1994 * , a49 , 535 . kozhekin a. ; kurizki g. ; sherman b. _ phys . rev . _ * 1996 * , a54 , 3535 . foster g.t . ; orozco l.a . ; castro - beltran h.m . ; carmichael h.j . _ _ * 2000 * , 85 , 3149 . gosh h. ; gerry c.c . * 1997 * , b14 , 2782 . * 2007 * , b24 , 2619 . valverde c.;baseia b. _ arxiv_:1609.01665v1 [ quant - ph ] . moya - cessa h. ; knight p.l . ; rosenhouse - dantsker a. _ phys . * 1994 * , 50 , 1814 . shore b.w . ; knight p.l . ; _ j. of mod . opt . _ * 1993 * , 40 , 1195 . wallentowitz s. ; vogel w. _ phys . * 1997 * , 55 , 4438 . moya - cessa h. ; soto - eguibar f. ; vargas - martinez j.m . ; juarez - amaro r.;a . zuiga - segundo _ physics reports _ * 2012 * , 513 , 229 - 261 . stenholm s. _ phys . rep . _ * 1973 * , 6 , 1 . _ z. phys . _ * 1927 * , 40 , 193 . susskind l. ; glogower j. _ physics _ * 1964 * , 1 , 49 . gerry c.c . ; knight p.l . _ introductory quantum optics . _ * 2005 * , cambridge university press . mandel l. _ physica scripta _ * 1986 * , t12 , 34 . loudon r. ; knight p.l . _ j. of mod . optics _ * 1987 * , 34 , 709 . kuklinski j. r. ; madajczyk j.l . * 1988 * , 37 , 3175 . satyanarayana m.v . ; rice p. ; vyas r. ; carmichael h.j . b * 1989 * , 6 , 228 . moya - cessa h. ; vidiella - barranco a. _ j. of mod . optics _ * 1992 * , 39 , 2481 .	we show that it is possible to add or subtract many photons from a cavity field by interacting it resonantly with a two - level atom . the atom , after entangling with the field inside the cavity and exiting it , may be measured in one of the schmidt states , producing a multiphoton process ( in the sense that can add or annihilate more photons than a single transition allows ) , i.e. , adding or subtracting several photons from the cavity field .
over the past century , researchers have devoted considerable effort into studying animal grouping behavior due to its important implications for social intelligence , collective cognition , and potential applications in engineering , artificial intelligence , and robotics @xcite . indeed , grouping behaviors are pervasive across all forms of life . for example , european starlings ( _ sturnus vulgaris _ ) are known to form murmurations of millions of birds which perform awe - inspiring displays of coordinated movement @xcite . western honeybees ( _ apis mellifera _ ) communicate the location of food and nest sites to other bees in their group via a complex dance language @xcite . even relatively simple bacteria exhibit grouping behavior , such as _ escherichia coli _ forming biofilms which allow their group to survive in hostile environments @xcite . _ swarming _ is one example of grouping behavior , where animals coordinate their movement with conspecifics to maintain a cohesive group . although swarm - like groups could arise by chance , e.g. , little egrets ( _ egretta garzetta _ ) pursuing a common resource in water pools @xcite , typically swarms are maintained via behavioral mechanisms that ensure group cohesion @xcite . as with many traits , swarming behavior entails a variety of fitness costs , such as increased risk of predation and the requisite sharing of resources with the group @xcite . with this fact in mind , significant effort has been dedicated to understanding the compensating benefits that grouping behavior provides @xcite . many such benefits of grouping behavior have been proposed , for example , swarming may improve mating success @xcite , increase foraging efficiency @xcite , or enable the group to solve problems that would be impossible to solve individually @xcite . furthermore , swarming behaviors are hypothesized to protect group members from predators in several ways . for example , swarming can improve group vigilance @xcite , reduce the chance of being encountered by predators @xcite , dilute an individual s risk of being attacked @xcite , enable an active defense against predators @xcite , or reduce predator attack efficiency by confusing the predator @xcite . unfortunately , many swarming animals take months or even years to produce offspring . these long generation times make it extremely difficult to experimentally determine which of the aforementioned benefits are sufficient to select for swarming behavior as an evolutionary response , let alone study the behaviors as they evolve @xcite . in this paper , we use a digital model of predator - prey coevolution to explore hamilton s selfish herd hypothesis @xcite . briefly , the selfish herd hypothesis states that prey in groups under attack from a predator will seek to place other prey in between themselves and the predator , thus maximizing their chance of survival . as a consequence of this selfish behavior , individuals continually move toward a central point in the group , which gives rise to the appearance of a cohesive swarm . in our model , both predators and prey have the ability to detect and interact with other agents in the environment . we evolve the agents with a genetic algorithm by preferentially selecting predators and prey based on how effective they are at consuming prey and surviving , respectively . forming a selfish herd is a possible solution for the prey to survive longer , but is not selected for directly . in this study , we first test whether a selfish herd evolves within a two - dimensional virtual environment with different forms of simulated predation . doing so enables us to experimentally control the specific modes of predation . we find that if predators are able to consistently attack the center of the group of prey , the selfish herd will not evolve . in subsequent experiments , we discover that density - dependent predation can provide a generalization of hamilton s original formulation of `` domains of danger . '' following these findings , we coevolve groups of predators and prey in a similar virtual environment to determine if coevolving predators impact the likelihood of the selfish herd to evolve . consequently , this study demonstrates that density - dependent predation provides a sufficient selective advantage for prey to evolve selfish herd behavior in response to predation by coevolving predators . finally , we analyze the evolved control algorithms of the swarming prey and identify simple , biologically - plausible agent - based algorithms that produce emergent swarming behavior . a preliminary investigation of the selfish herd hypothesis was presented at the genetic and evolutionary computation conference in 2013 @xcite . this paper expands on that work by studying the long - term evolutionary effects of differing attack modes , exploring a new attack mode that directly selects against selfish herd behavior , and providing an analysis of the control algorithms that evolved in the swarming prey . hamilton s original formulation of the selfish herd hypothesis introduced the concept of `` domains of danger '' ( dods , figure [ fig : domains - of - danger ] ) , which served as a method to visualize the likelihood of a prey inside a group to be attacked by a predator @xcite . prey on the edges of the group would have larger dods than prey on the inside of the group ; thus , prey on the edges of the group would be attacked more frequently . moreover , hamilton proposed that prey on the edges of the group would seek to reduce their dod by moving inside the group , thus placing other group members between themselves and the predator . further work has expanded on this hypothesis by adding a limited predator attack range @xcite , investigating the effects of prey vigilance @xcite , considering the initial spatial positioning of prey when the group is attacked @xcite , and even confirming hamilton s predictions in biological systems @xcite . additional studies have focused on the movement rules that prey in a selfish herd follow to minimize their dod @xcite . this line of work began by demonstrating that the simple movement rules proposed by hamilton reduce predation for prey inside the group @xcite , then opened some parameters of the movement rules to evolution in an attempt to discover a more biologically plausible set of movement rules @xcite . finally , some studies have investigated the evolution of predator behavior in response to prey density @xcite , the coevolution of predator and prey behavior in the presence of the predator confusion effect @xcite , and elaborating upon the interaction between ecology and the evolution of grouping behavior @xcite . this paper builds on this work by studying the effects of coevolving predators and predator attack mode ( i.e. , how predators select a prey in a group to attack ) on the evolution of the selfish herd . more broadly , in the past decade researchers have focused on the application of locally - interacting swarming agents to optimization problems , called particle swarm optimization ( pso ) @xcite . pso applications range from feature selection for classifiers @xcite , to video processing @xcite , to open vehicle routing @xcite . a related technique within pso seeks to combine pso with coevolving `` predator '' and `` prey '' solutions to avoid local minima @xcite . researchers have even sought to harness the collective problem solving power of swarming agents to design robust autonomous robotic swarms @xcite . thus , elaborations on the foundations of animal grouping behavior has the potential to improve our ability to solve engineering problems . to study the evolution of the selfish herd , we developed an agent - based model in which agents interact in a continuous , toroidal virtual environment ( @xmath0 virtual meters ) , shown in figure [ fig : sim - env ] . at the beginning of each simulation , we place 250 agents in the environment at uniformly random locations . these agents are treated as `` virtual prey . '' each agent is controlled by a _ markov network _ ( mn ) , which is a probabilistic controller that makes movement decisions based on a combination of sensory input ( i.e. , vision ) and internal states ( i.e. , memory ) @xcite . we evolve the agent mns with a genetic algorithm ( ga ) @xcite under varying selection regimes , which will be described in more detail below . during each simulation time step , all agents read information from their sensors and take action ( i.e. , move ) based on their effectors . in our first set of treatments , we simulate an ideal , disembodied predator by periodically removing prey agents from the environment and marking them as consumed , e.g. , when they are on the outermost edges of the group . subsequent treatments introduce an embodied , coevolving predator agent which is controlled by its own mn . the source code and data for these experiments are available online . in the remainder of this section , we describe the sensory - motor architecture of individual agents and present details related to the function and encoding of mns . limited - distance retina ( 200 virtual meters ) to observe their surroundings and detect the presence of other agents . the current heading of the agent is indicated by a bold arrow . each agent has its own markov network , which decides where to move next based off of a combination of sensory input and memory . the left and right actuators ( labeled `` l '' and `` r '' ) enable the agents to move forward , left , and right in discrete steps.,scaledwidth=55.0% ] figure [ fig : agent - illustration ] depicts the sensory - motor architecture of the agents used for this study . a prey agent can sense predators and conspecifics with a limited - distance ( 200 virtual meters ) , pixelated retina covering its entire 360@xmath1 visual field . its retina is split into 24 even slices , each covering an arc of 15@xmath1 , which is an abstraction of the broad , coarse visual systems often observed in grouping prey @xcite . regardless of the number of agents present in a single retina slice , the prey agent only knows whether a conspecific or predator resides within that slice , but not how many . for example , in figure [ fig : agent - illustration ] , the fourth retina slice to the right of the agent s heading ( labeled `` a '' ) has both the predator and prey sensors activated because there are two predator agents and a prey agent inside that slice . once provided with its sensory information , the prey agent chooses one of four discrete actions , as shown in table [ table : agent - action - encoding ] . prey agents turn in 8@xmath1 increments and move 1 virtual meter each time step . in our coevolution experiments , the predator agents detect nearby prey agents and conspecifics using a limited - distance ( 200 virtual meters ) , pixelated retina covering its frontal 180@xmath1 that works just like the prey agent s retina ( figure [ fig : agent - illustration ] ) . similar to the prey agents , predators make decisions about how to move next using their mn , as shown in table [ table : agent - action - encoding ] , but move 3@xmath2 faster than the prey agents and turn correspondingly slower ( 6@xmath1 per simulation time step ) due to their higher speed . finally , if a predator agent moves within 5 virtual meters of a prey agent that is anywhere within its retina , the predator agent makes an attack attempt on the prey agent . if the attack attempt is successful , we remove the prey agent from the simulation and mark it as consumed . each agent is controlled by its own markov network ( mn ) , which is a probabilistic controller that makes decisions about how the agent interacts with the environment and other agents within that environment . since a mn is responsible for the control decisions of its agent , it can be thought of as an _ artificial brain _ for the agent it controls . every simulation time step , the mns receive input via sensors ( e.g. , visual retina ) , perform a computation on inputs and any hidden states ( i.e. , memory ) , then place the result of the computation into hidden or output states ( e.g. , actuators ) . we note that mn states are binary and only assume a value of 0 or 1 . when we evolve mns with a ga , mutations affect ( 1 ) which states the mn pays attention to as input , ( 2 ) which states the mn outputs the result of its computation to , and ( 3 ) the internal logic that converts the input into the corresponding output . .possible actions encoded by the agent s output . each output pair encodes a discrete action taken by the agent . the agent s mn changes the values stored in output states l and r to indicate the action it has decided to take in the next simulation time step . [ cols="<,<,<",options="header " , ] [ table : hdaa - treatments ] thus far , we have explored attack modes that select for the evolution of swarming behavior . it is not surprising that there are also attack modes exhibited by natural predators that must select against swarming behavior in their prey . for example , blue whales ( _ balaenoptera musculus _ ) are known to dive into the densest areas in swarms of krill , consuming hundreds of thousands of krill in the middle of the swarm in a single attack @xcite . we call this kind of attack mode a _ high - density area attack_. such an attack clearly selects against swarming behavior because it targets the prey that swarm the most . if krill swarms consistently experience these high - density area attacks , then why do they still evolve swarming behavior ? it is important to note that krill swarms are also fed on by smaller species , such as crabeater seals ( _ lobodon carcinophagus _ ) , that consistently attack the krill on the outside of the swarm @xcite . thus , krill swarms are experiencing two forms of attack modes simultaneously : high - density area attacks from whales and outside attacks from crabeater seals . thus , it is possible that the selection pressure to swarm from outside attacks ( figure [ fig : sdc - artificial - selection ] ) could outweigh the selection pressure to disperse from high - density area attacks . shown in figure [ fig : attack - modes]d , we model high - density area attacks as an artificial attack that always targets the prey at the most dense area of the swarm ( i.e. , highest @xmath3 ) . we note that this attack mode is the opposite of the density - dependent mechanism explored in the previous section , which favors predators that target prey in the _ least _ dense area of the swarm . once the target is selected , we execute the attack by removing the target prey and all other prey within 30 virtual meters of the target prey . outside attacks are modeled as described above . to study the effect of high - density area attacks on the evolution of swarming behavior , we allow the prey to evolve while experiencing both attack modes simultaneously . we vary the relative handling times of both attacks ( table [ table : hdaa - treatments ] ) to explore whether relative attack frequency could explain why some swarming animals evolved swarming behavior despite the fact that they experience high - density area attacks . as shown in figure [ fig : sdc - artificial - selection - hdaa ] , prey experiencing only outside attacks quickly evolve cohesive swarming behavior ( light grey triangles with a full line ) . however , when we introduce infrequent high - density area attacks ( dark grey circles with a dashed line ) , the selection pressure for prey to swarm is reduced . finally , when we introduce frequent high - density area attacks ( black squares with a dotted line ) , the prey do not evolve swarming behavior at all . thus , one possible explanation for animals evolving swarming behavior despite experiencing high - density area attacks is that the high - density area attacks are too infrequent relative to other attack types to exert a strong enough selection pressure for prey to disperse . in summary , the artificial selection experiments provided us with two important pieces of information regarding the evolution of the selfish herd : ( 1 ) attacks on prey on the periphery of the herd exert a strong selection pressure for prey to swarm and ( 2 ) prey in less dense areas , such as those on the outside of the herd , must experience a higher predation rate than in areas of dense prey , such as found in the center of the herd . building upon the artificial selection experiments , we implemented density - dependent predation in a predator - prey coevolution experiment . adding predators into the simulation environment enables us to observe how embodied coevolving predators affect the evolution of the selfish herd . for this experiment , we coevolve a population of 100 predator genomes with a population of 100 prey genomes using a ga with settings described in table [ table : ga - settings ] . specifically , we evaluate each predator genome against the entire prey genome population for 2,000 simulation time steps each generation . during evaluation , we place 4 clonal predator agents inside a @xmath4 virtual meters simulation environment with all 100 prey agents and allow the predator agents to make attack attempts on the prey agents . the prey genome population size , simulation environment area , and total number of ga generations were decreased in this experiment due to computational limitations imposed by predator - prey coevolution . we assigned the prey individual fitness values as in the previous experiments , and evaluated predator fitness according to the following equation : @xmath5 where @xmath6 is the current simulation time step , @xmath7 is the total number of simulation time steps ( here , @xmath7 = 2,000 ) , @xmath8 is the starting group size ( here , @xmath8 = 100 ) , and @xmath9 is the number of prey alive at update @xmath6 . thus , predators are selected to consume more prey faster , and prey are selected to survive longer than other prey in the group . once all of the predator and prey genomes are assigned fitness values , we perform fitness proportionate selection on the populations via a moran process @xcite , increment the generation counter , and repeat the evaluation process on the new populations until the final generation ( 1,200 ) is reached . to evaluate the coevolved predators and prey quantitatively , we obtained the line of descent ( lod ) for every replicate by tracing the ancestors of the most - fit prey mn in the final population until we reached the randomly - generated ancestral mn with which the starting population was seeded ( see @xcite for an introduction to the concept of a lod in the context of digital evolution ) . we again characterized the prey grouping behavior by measuring the swarm density of the entire prey population every generation . figure [ fig : sdc - pred - prey - coevolution ] depicts the prey behavior measurements for the coevolution experiments with density - dependent predation ( black circles with a dashed line ; mean swarm density at generation 1,200 @xmath10 two standard errors : [email protected] ) and without density - dependent predation ( light grey triangles with a full line ; [email protected] ) . without density - dependent predation , the prey evolved purely dispersive behavior as a mechanism to escape the predators . however , with density - dependent predation , the prey evolved cohesive swarming behavior in response to attacks from the predators . here we see that density - dependent predation provides a sufficient selective advantage for prey to evolve the selfish herd in response to predation by coevolving predators . accordingly , these results uphold hamilton s hypothesis that grouping behavior could evolve in animals purely due to selfish reasons , without the need for an explanation that involves the benefits to the whole group @xcite . moreover , the discoveries in this work refine the selfish herd hypothesis by clarifying the effect that different attack modes have on the evolution of the selfish herd . now that we have evolved emergent swarming behavior in an agent - based model under several different treatments , we can analyze the resulting markov networks ( mns ) to gain a deeper understanding of the individual - based mechanisms underlying swarming behavior . for this analysis , we chose the most - abundant prey mn from each of the outside attack artificial selection experiment replicates , resulting in 100 mns that exhibit swarming behavior . first , we analyze the structure of the 100 mns by looking at the specific retina sensors that the mns evolved to connect to . shown in figure [ fig : prey - brain - connectivity ] , the prey mns show a strong bias for connecting to the prey - specific retina sensors in front of the prey , but not to the sides . additionally , some of the prey mns show a preference for connecting to the prey - specific retina sensors behind the prey . from this analysis alone , we can deduce that the retina sensors that are most conducive to swarming behavior are in front of the prey agent . to understand how prey make movement decisions based on their sensory input , we map every possible input combination in the prey s retina to the corresponding movement decision that the prey made . due to the stochastic nature of markov networks , the prey agents do not always make the same movement decision when given the same input . thus , we take the most - likely output out of 1,000 trials as the representative decision for a given sensory input combination . effectively , this process produces a truth table that maps every possible sensory input to its corresponding movement decision . we input this truth table into the logic minimization software _ espresso _ @xcite , which outputs the minimal representative logic of the truth table . this process results in a truth table that is reduced enough to make the evolved prey behavior comprehensible by humans . surprisingly , the individual - based mechanisms underlying the emergent swarming behavior are remarkably simple . most of the prey mns evolved to make their movement decisions based off of only one prey sensor in front of the prey agent . if the prey sensor does not detect another prey agent , the agent repeatedly turns in one direction until it detects another prey agent in that sensor . once the agent detects another prey agent in the sensor , it moves forward until the agent is no longer visible . this mechanism alone proved sufficient to produce cohesive swarming behavior in the majority of our experiments . interestingly , this discovery corroborates the findings in earlier studies suggesting that complex swarming behavior can emerge from simple movement rules when applied over a population of locally - interacting agents @xcite . in a small subset of the evolved prey mns , we observe mns that occasionally connect to one of the prey sensors behind them . these mns watch for a prey agent to appear in a single prey sensor behind the agent and turn repeatedly in one direction until a prey agent is no longer visible in that sensor . once a prey agent is no longer visible in the back sensor , the mn moves forward or turns depending on the state of the frontal sensor . we note that this mechanism _ only _ evolved in prey mns that already exhibited swarming behavior using one of the frontal sensors , which suggests that this mechanism does not play a major role in swarming behavior . instead , this mechanism seems to cause the prey agent to turn toward the center of the swarm instead of swarming in a circle with the rest of the prey agents . this mechanism can be thought of as a `` selfish herd '' mechanism that attempts to selfishly move the agent toward the center of the swarm to avoid predation . the contributions of this work are as follows . first , we demonstrate hamilton s selfish herd hypothesis in a digital evolutionary model and highlight that it is the attack mode of the predator that critically determines the evolvability of selfish herd behavior . second , we show that density - dependent predation is sufficient for the selfish herd to evolve as long as the predators can not consistently attack prey in the center of the group . finally , we show that density - dependent predation is sufficient to evolve grouping behavior in prey as a response to predation by coevolving predators . consequently , future work exploring the evolution of the selfish herd in animals should not only consider the behavior of the prey in the group , but the attack mode of the predators as well . following these experiments , we analyze the evolved control algorithms of the swarming prey and identify simple , biologically - plausible agent - based algorithms that produce emergent swarming behavior . while this work shows one method by which the the evolution of grouping behavior can be studied , there remain many different hypotheses explaining the evolution of grouping behavior @xcite . our future work in this area will focus on directly testing these hypotheses in similar digital evolutionary models . this research has been supported in part by the national science foundation ( nsf ) beacon center under cooperative agreement dbi-0939454 , and nsf grant oci-1122617 . any opinions , findings , and conclusions or recommendations expressed in this material are those of the author(s ) and do not necessarily reflect the views of the nsf . we gratefully acknowledge the support of the michigan state university high performance computing center and the institute for cyber enabled research ( icer ) . olson , d.b . knoester , and c. adami . critical interplay between density - dependent predation and evolution of the selfish herd . in _ proceedings of the genetic and evolutionary computation conference ( gecco ) _ , pages 247254 , 2013 . h. kunz , t. zblin , and c.k . hemelrijk . on prey grouping and predator confusion in artificial fish schools . in _ proceedings of the international conference on the simulation and synthesis of living systems ( alife ) _ , pages 365371 , 2006 . l. spector , j. klein , c. perry , and m. feinstein . emergence of collective behavior in evolving populations of flying agents . in _ proceedings of the genetic and evolutionary computation conference ( gecco ) _ , pages 6173 , 2003 . b. xue , m. zhang , and w.n . multi - objective particle swarm optimisation ( pso ) for feature selection . in _ proceedings of the genetic and evolutionary computation conference ( gecco ) _ , pages 8188 , 2012 . e. vellasques , r. sabourin , and e. granger . gaussian mixture modeling for dynamic particle swarm optimization of recurrent problems . in _ proceedings of the genetic and evolutionary computation conference ( gecco ) _ , pages 7380 , 2012 . y. marinakis and m. marinaki . a honey bees mating optimization algorithm for the open vehicle routing problem . in _ proceedings of the genetic and evolutionary computation conference ( gecco ) _ , pages 101108 , 2011 . a. silva , a. neves , and e. costa . an empirical comparison of particle swarm and predator prey optimisation . in _ proceedings of the irish conference on artificial intelligence and cognitive science ( aics ) _ , pages 103110 , 2002 . goldbogen , j. calambokidis , e. oleson , j. potvin , n.d . pyenson , g. schorr , and r.e . mechanics , hydrodynamics and energetics of blue whale lunge feeding : efficiency dependence on krill density . , 214:131146 , 2011 .	animal grouping behaviors have been widely studied due to their implications for understanding social intelligence , collective cognition , and potential applications in engineering , artificial intelligence , and robotics . an important biological aspect of these studies is discerning which selection pressures favor the evolution of grouping behavior . in the past decade , researchers have begun using evolutionary computation to study the evolutionary effects of these selection pressures in predator - prey models . the selfish herd hypothesis states that concentrated groups arise because prey selfishly attempt to place their conspecifics between themselves and the predator , thus causing an endless cycle of movement toward the center of the group . using an evolutionary model of a predator - prey system , we show that how predators attack is critical to the evolution of the selfish herd . following this discovery , we show that density - dependent predation provides an abstraction of hamilton s original formulation of `` domains of danger . '' finally , we verify that density - dependent predation provides a sufficient selective advantage for prey to evolve the selfish herd in response to predation by coevolving predators . thus , our work corroborates hamilton s selfish herd hypothesis in a digital evolutionary model , refines the assumptions of the selfish herd hypothesis , and generalizes the domain of danger concept to density - dependent predation . group behavior , selfish herd theory , predator attack mode , density - dependent predation , predator - prey coevolution , evolutionary algorithm , digital evolutionary model .
since the advent of the copenhagen interpretation of quantum mechanics@xcite , there have been many distinct proposals from quantum physicists with various backgrounds to resolve the problems of measurement and the origin of the quantum state reduction that generates non - unitary time evolution of state vectors@xcite . in this paper , as a new approach , we address a connection between time itself and non - unitary time evolution in quantum mechanics by extending canonical gravity theory . our mathematical model is based on the reformulation of canonical gravity as a _ spin system _ by amitabha sen@xcite and part of our approach uses the notion of the penrose uncertainties assigned to superposed space - times@xcite which we relate to the quantum superposition of originally definite particle positions or momenta in a coherent second - quantized system . an original aspect of our approach is that we extend sen s spin model by introducing in the low - velocity approximation one additional long - range self - interacting massive spinor particle that is physically coupled with sen s spinor . in this paper , we call it the _ theta particle_. in the present model , we assume that the role of theta particle is such that this particle only interacts with time lapses and the causal structure of the world , that is , the temporal part of the space - time metric , via its coupling to sen s spinor and has _ no _ direct interaction with the standard model elementary particles except for the origin of its mass . our strategy for deriving non - unitary time evolution in the world , as measured by its proper time , _ of _ a spatially macroscopically coherent quantum matter system by solving the schr@xmath0dinger equation written using time increments is to show that fluctuations of the time increment around its trial mean are induced by a _ phase transition _ , that is , _ spontaneous _ time reparametrization symmetry ( trps ) breaking refers to that of the global symmetry part ( i.e. , the pre - gauged symmetry ) of @xmath1 . thus , the corresponding _ ground state expectation values _ are functions of the space - time coordinates . ] , where the theta particle plays the role of creating the trp invariant potential . in canonical quantum gravity , trps is manifest in the wheeler - de witt equation of the wave function of the universe and its reduced form in matter systems where there is no newtonian concept of external time and only an internal clock exists@xcite . such a situation is caused by the fact that , besides canonical gravity being a totally constrained system , the object of its dynamics is the universe itself . in the quantum mechanical regime , our formalism employs a time increment rather than a time parameter in order to treat the statistics of time increments . in the trps broken phase , the time increment trials have non - zero statistical variance . the ratio between the variance and the mean of the time increment trials , denoted by @xmath2 , is the scale of trps breaking with dimensions of time . the plan of this paper is as follows . in sec . ii we present the model hamiltonian of spinor gravity coupled to the theta particle based on sen s reformulation of canonical gravity as a spin system . in sec . iii we describe the mechanism of trps breaking and the consequent non - unitary evolution of quantum systems in our model . in sec . iv we summarize our results and remark on some interesting features of our model . in this section , we construct our model . to do this , we start from the reformulation of canonical gravity by sen , keeping its generality.@xcite in this paper , to study the time evolution of space - time structure , we use the arnowitt - deser - misner ( adm ) @xmath3 decomposition of a space - time @xmath4@xcite in the canonical formalism . in the time evolution of a space - like hypersurface @xmath5 in @xmath6 , the initial data are a pair of symmetric tensor fields @xmath7 , that is , the metric @xmath8 ( for which we choose the signature @xmath9 ) induced from @xmath4 and the second fundamental form @xmath10 . in the adm decomposition method , the square of the line element in @xmath6 is written as @xmath11 for time lapse function @xmath12 and shift vector @xmath13 . here , we clarify two points concerning eq.([eq : adm]).@xcite first , when we regard @xmath5 as the embedding of a space manifold @xmath14 into space - time @xmath6 , denoted by @xmath15 , the lapse function @xmath16 and the shift vector @xmath17 are the components of the deformation of the embedding @xmath18 perpendicular and parallel to @xmath5 , respectively . second , due to their role as lagrange multipliers for the @xmath3 decomposition in the totally constrained system , the lapse function @xmath12 and the shift vector @xmath13 are the time - dependent variables of trps and the space diffeomorphism of @xmath14 , respectively . the arbitrariness of these four variables in the space - time metric @xmath19 reflects that of a choice of space - time coordinate system . now , following ref.@xcite , we consider an @xmath20 spinor field variable @xmath21 on @xmath22 and restrict it to @xmath5 . we call this the _ space spinor_. we denote by @xmath23 the field of unit vectors on @xmath22 that is normal to the family of @xmath5 in the @xmath3 decomposition . we restrict @xmath23 to @xmath5 and then represent the covariant four - vector @xmath24 , at each point @xmath25 in @xmath5 , by a natural hermitian isomorphism @xmath26 between the complex conjugate space , @xmath27 , of the complex two - dimensional vector space @xmath28 of primed space spinors of arbitrarily given kind at @xmath25 and the dual space , @xmath29 , of the vector space @xmath30 of unprimed space spinors of one kind : @xmath31 here , we lower and raise space spinor indices by contraction with a non - degenerate symplectic form @xmath32 , with which we equip @xmath30 , having a skew pair of space spinor indices . this isomorphism @xmath26 , as an additional structure on @xmath30 , reduces the structure group acting on @xmath30 at each point @xmath25 from @xmath20 to @xmath33 . so , space spinors obtained from elements of @xmath30 by varying @xmath25 over @xmath5 are in fact @xmath33 spinors on @xmath5 . due to this feature of space spinors , we can relate them to the geometry of @xmath5 with the induced negative - definite metric @xmath34 such that @xmath35 holds . for instance , we can express _ spatial _ tensors on @xmath5 that vanish by contraction with @xmath23 or @xmath36 in terms of space spinors by replacing each spatial tensor index by a symmetric pair of space spinor indices : for example , we write @xmath37 as @xmath38 such that @xmath39 holds for the @xmath33 adjoint @xmath40 , since @xmath41 holds due to the unity of @xmath23 . to reformulate canonical gravity in terms of space spinors , sen introduced @xmath33 connections @xmath42 by @xmath43 where @xmath44 is the derivative operator on @xmath45 and @xmath21 is a space spinor on @xmath5.@xcite moreover , sen modeled the lapse function @xmath12 and the shift vector @xmath46 as two kinds of squares of a space spinor @xmath21@xcite : @xmath47 as a result , time lapses and the causal structure of the world are encoded in sen spinor field @xmath21 . the total hamiltonian @xmath48 of our model consists of four parts : @xmath49 the proper time @xmath50 of @xmath45 appears in the combination @xmath51 . in units where @xmath52 and @xmath53 , sen rewrote the hamiltonian of gravity in the space spinor formalism as@xcite @xmath54ds^{(\alpha\beta)(\mu\nu)(\lambda\rho)}\;.\label{eq : sen}\end{aligned}\ ] ] integrating by parts and using eq.([eq : nn ] ) shows that this hamiltonian is equivalent to the adm hamiltonian with a boundary term @xmath55 giving the adm mass@xcite : @xmath56 where @xmath57 both of which must vanish ( see , for example , ref.@xcite ) by including the contributions from other parts @xmath58 and @xmath59 as constraints with lagrange multipliers @xmath12 and @xmath13 , respectively , in the totally constrained system @xmath60 for the scalar curvature @xmath61 on @xmath45 and @xmath62 . in this paper , our interest lies in the spatial weak - field case @xmath63 now , we are ready to introduce the spin @xmath64 fermionic theta particle field on @xmath5 into our model and extend canonical gravity . we do this in a way that its four - dimensional spin - variable space matches that of sen spinor field : it is the composite of _ its own _ @xmath21 and a scalar particle . at this stage it is necessary to make a technical point . in this study , we hypothesize theta particles to be _ cold dark matter _ and to simplify the model setting , we deal with only theta particles with velocities that are low relative to the speed of light . then , the theta particle field is approximately described by a non - relativistic pauli @xmath33 two - component spinor field , @xmath65 , on @xmath5 . in the @xmath66 vector spin model , in which all variables are treated as their field amounts , we define the unit modulus and the direction of theta spin by @xmath67 and require that it physically couples to sen spin @xmath21 by two different kinds of strong coupling ( @xmath68 ) : @xmath69 the two terms of @xmath70 are separately invariant under trp and spatial diffeomorphisms . our system @xmath71 gives energy to sen spin only by its couplings to a theta particle in @xmath70 . this form of energy enables a certain kind of locally isolated theta particle system to generate a measuring system in quantum mechanics . here , we make two remarks about eq.([eq : hc ] ) . first , these physical couplings mean that , in our model , the lapse function and the shift vector play gauge independent roles . second , the first term of eq.([eq : hc ] ) is regarded as a spin - exchange interaction between two kinds of sen spins with common spin eigenspaces . a spin - exchange interaction is the reduced form of an orbital interaction that is obtained by suppressing orbital degrees of freedom . next , we model the hamiltonian of the theta particle field @xmath65 itself as that of a spin model with a long - range ferromagnetic interaction via a massless wave of a continuous succession of spin - exchange interactions , which propagates in the background @xmath21 _ of _ @xmath65 , @xmath72 where in the newtonian limit , @xmath73 for a theta particle field @xmath74 with mass @xmath75 , canonical momentum @xmath76 and a positive - valued long - range decay factor @xmath77 , with a constant @xmath78 , that generates an inverse square law force . the hamiltonian @xmath79 of the matter fields @xmath80 is the matter sector of general relativity@xcite , which consists of the coupling of the lapse function @xmath12 to the component @xmath81 and the coupling of the shift vector @xmath13 to the component @xmath82 as in @xmath83 . in the form presented here , our system @xmath48 manifests a global @xmath33 spin rotation symmetry . this global symmetry is gauged by the background @xmath21 _ of _ @xmath65 . the chain of interactions and correlations between @xmath84 , @xmath85 and @xmath80 is schematically shown in the diagram @xmath86\|{(a)}\ar[r]^{(b)}&\ar[l]&\lambda\ar@(ul , dl)[]\|{(c)}\ar@/_0pt/@{.>}[r]^{(d)}&\psi\ar@/_0pt/@{.>}[l]\ar@/_25pt/@{.>}[lll]_{(e)}}\;.\label{eq : int}\ ] ] in our model , while @xmath87 is a short - range interaction , @xmath88 is a long - range self - interaction . the process @xmath89 defines the lapse function @xmath12 and the shift vector @xmath13 , and the right - directed arrow of @xmath90 arises from time - dependent processes in the world for @xmath80 with its proper time including the adm time lapse . [ for the derivation of the matter schr@xmath0dinger equation from the wheeler - de witt equation in the wkb approximation , see ref.@xcite . ] the correlation @xmath91 is indirect . as will be remarked at the end of this paper , @xmath91 is causally a _ one - way _ process . this paper s main statement is that , in our model , trps can be broken in time - dependent processes in the world for a spatially macroscopically coherent quantum matter system @xmath80 and this trps breaking leads to non - unitary time evolution in this world . throughout this paper , the term _ coherence _ is used with the meaning of quantum coherence of superposed de broglie waves and superposed photons@xcite . in this section , we describe the mechanism behind our main statement based on the diagram shown as eq.([eq : int ] ) . in the following , we denote the first and the second terms of @xmath70 by @xmath92 and @xmath93 , respectively . we assume a second - quantized matter system @xmath80 of @xmath95-number of identical particles , with the coherence of @xmath80 limited by macroscopic lengths @xmath96 ( @xmath97 ) at a time @xmath98 . then , the existence of the macroscopically , in the quantum mechanical sense , ferromagnetically ordered distribution of @xmath95-number of sen spins over the time - dependent _ coherence domain _ of @xmath80 in @xmath5 at the time @xmath98 follows . ( the @xmath95-number of sen spins stochastically accompany a second - quantized @xmath95-body system @xmath80 comprising arbitrary kinds of particles , with the same probability as @xmath80 , while the position uncertainty of @xmath80 is kept . ) the key fact for showing this existence is that the dimensions @xmath96 of the coherence domain are smaller than the spatial stochastic means of the penrose position uncertainties @xmath99 , in the identification between the space - times having the probability amplitudes of @xmath80 under superposition , taken as the root mean square error ( see below ) in spatial identification at each point @xmath100 in the time slice @xmath5 for @xmath97 , respectively : @xmath101 for a field of unit vectors @xmath102 , the root mean squares @xmath103 weighted by the one - particle probability density @xmath104 of @xmath80 in the particle picture as in @xmath105 itself@xcite , the vector modulus of the expectation values @xmath106 , and the mandelstam - tamm time interval @xmath107@xcite characteristic of the rate of change of the system @xmath45 itself at the time @xmath98 . ( note three points : @xmath105 can be determined only stochastically , each @xmath108 is not a canonical variable of @xmath109 , and @xmath21 has no orbital degree of freedom . ) in eq.([eq : dpo ] ) , @xmath110 is the direction of the group velocity @xmath111 of the displacement of @xmath5 in @xmath22 . this direction is arbitrarily chosen , due to the @xmath33 gauge symmetry , and needs to be definite within the coherence domain that is analogous to a moving wave packet . when we define the _ magnetization _ @xmath22 of a system of uniformly distributed spins by the vector modulus of ( weighted ) arithmetic means of the spin vector components over this system , eq.([eq : dpo ] ) shows that approximately a finite magnetization , weighted by @xmath112 , of @xmath95-number of expected sen spatial spin vectors of the order of @xmath113 is generated for the root mean square @xmath114 weighted by @xmath112 . when @xmath95 is large enough , the configuration of added sen spins generates a magnetized ( i.e. , ferromagnetically ordered ) distribution of spin vectors of theta particles @xmath115 via the strong ferromagnetic coupling @xmath116 . this effect induces the spontaneous breaking of the @xmath33 gauge symmetry of the total spin system within the coherence domain . here , an important point is that , due to eq.([eq : xy ] ) , interactions between these aligned theta spins in the magnetized system are attractive . such that @xmath117 is the trps breaking scale ( i.e. , the fluctuation scale of the proper time ) @xmath2 and above . then , spatial diffeomorphism invariance within the coherence domain would be broken spontaneously . ( by replacing @xmath118 with @xmath119 , the same line of argument for symmetry breaking as in step c will show that the dilation symmetry on the shift vector would be broken spontaneously in the total spin system . ) ] once such a spatially macroscopically coherent quantum system accompanies a locally isolated theta particle system with aligned spins , the theta particles in it settle down to their local collisionless equilibrium state due to the long - range nature of the attractive interactions.@xcite based on recent results in the mean - field analysis of the three - dimensional self - gravitating model and the simplest benchmark model for systems of the type given in eq.([eq : xy ] ) ( see figs . 1 and 2 ) and the universality of lynden - bell statistics in long - range interacting systems , we arguably infer that in general , the collisionless equilibrium one - particle state @xmath121 of this system , in the mean - field treatment , is realized as a superposition of two independent lynden - bell coarse - grained distributions @xmath122 and @xmath123 , called the _ double lynden - bell distribution _ in ref.@xcite , with two different chemical potentials.@xcite ( the evaporated masses@xcite , that is , the masses having completely escaped from the main cluster are excluded from the local system . ) here , due to the pauli exclusion principle , which is applicable because of the fermionic nature of @xmath65 having orbital degrees of freedom , and the alignment of the initial theta spins , we consider an initial phase space ( @xmath124-space ) distribution of the _ water - bag _ type ( i.e. , with a single non - zero fine - grained level ) . consists of the following assumptions : the distinguishability of particles , an exclusion rule , conservation of mass and energy , and an initial phase space distribution of the water - bag type.@xcite the exclusion rule , which is reflected in the degeneracy in the lynden - bell distributions in fig . 1 , is a dynamical effect that is the result of vlasov incompressibility . ] the robustness of these chemical potentials will be discussed at the end of this paper . the ground state of the theta spin system is @xmath121 at zero temperature ( i.e. , @xmath125 for @xmath126 ) . spin model defined at a point @xmath127 ( called the _ hamiltonian mean - field model _ ) is a widely studied benchmark model for long - range interacting systems.@xcite a marked double lynden - bell one - particle distribution function @xmath128 of one - particle energy @xmath129 in this model is shown as the collisionless equilibrium under incomplete violent relaxation.@xcite the total mass and the total energy of the system are conserved . the green ( upper ) and cyan ( lower ) domains divided by a thick curve are independent lynden - bell distribution functions @xmath130 and @xmath131 . the dashed gray curve is the lynden - bell equilibrium of the system . ] spin ground state are shown . each distribution is uniform and each @xmath132 spin has unit mass . in the mean - field treatment , the phase space variables are the spin direction @xmath133 of the @xmath132 spin located at the point @xmath127 and its conjugate momentum @xmath134 for time @xmath135 . there is the relation @xmath136 where @xmath137 is the stationary modulus of the magnetization of the system . the green ( left ) and cyan ( right ) domains correspond to those shown in fig . 1 . the gray curve in each graph is the edge of the lynden - bell equilibrium of the system at zero temperature.,title="fig : " ] spin ground state are shown . each distribution is uniform and each @xmath132 spin has unit mass . in the mean - field treatment , the phase space variables are the spin direction @xmath133 of the @xmath132 spin located at the point @xmath127 and its conjugate momentum @xmath134 for time @xmath135 . there is the relation @xmath136 where @xmath137 is the stationary modulus of the magnetization of the system . the green ( left ) and cyan ( right ) domains correspond to those shown in fig . 1 . the gray curve in each graph is the edge of the lynden - bell equilibrium of the system at zero temperature.,title="fig : " ] here , two remarks on the _ collisionless equilibrium _ are made . first , in the collisionless equilibrium of a long - range interacting system , the variable of the fine - grained vlasov distribution is the one - particle energy only.@xcite the second remark concerns the persistence of this equilibrium . a general property of long - range interacting systems means that , in the thermodynamic limit , the theta particle system never approaches the boltzmann - gibbs equilibrium from the collisionless equilibrium and its kinetics is fully governed by the vlasov equation when we discuss nonequilibrium properties.@xcite in this step , we show that the trps breaking arises from the fact that , in the theta particle ground state , there are two _ coexisting _ different time lapse structures , coming from @xmath139 , with respect to the two lynden - bell energy distributions @xmath140 ( @xmath126 ) at zero temperature@xcite . we begin by noting that , in the total spin system , the theta particle that is coupled with the sen spin at a point @xmath100 on @xmath5 changes with time . this is because , while the sen spin is localized at every point in @xmath5 , every theta particle has momentum . in the collisionless equilibrium of the theta particle system , this itinerant property is reduced to the stationary distributions @xmath140 for @xmath126 . in the following , we denote by @xmath141 for @xmath126 the integrals over the momentum and the spin parts of @xmath140 with the spin direction dependence fixed by @xmath142 in step a. in the theta particle ground state , @xmath139 can be written as @xmath143 where @xmath144 is the theta particle number of the system , @xmath145 counts the couplings between the localized sen spin at a point @xmath100 on @xmath5 and an itinerant theta spin in @xmath146 , and the form of each coarse - grained position distribution @xmath141 is determined by the fermi energy of the lynden - bell distribution @xmath140 and the form of the self - consistent mean - field potential of the second term of eq.([eq : xy ] ) ( refer to fig . 2 ) . here , @xmath147 is the trp invariant potential for time lapse @xmath12 . the main property of this potential is that it gives distinct energetic degeneracies , with respect to trp , for two different time lapse densities @xmath148 for @xmath126 . owing to their differing trp degeneracies , the two time lapse densities define different proper times @xmath149 of @xmath140 and corresponding time lapses @xmath150 . this property of @xmath147 is confirmed by noting that , in the particle description , @xmath147 is written as @xmath151 where time lapses @xmath152 correspond to the time lapse densities @xmath148 in eq.([eq : pot ] ) . here , we note that this coarse - graining is applied not to the proper time @xmath135 but to the energy @xmath147 , so @xmath147 is necessary and @xmath153 needs to be non - zero . due to this property of @xmath147 , the ground state of the total spin system is in the form @xmath154 this expression means that for @xmath126 , @xmath155 where @xmath156 is the fraction of theta particles in @xmath146 . in eq.([eq : ti ] ) , for @xmath126 , the first and the second equalities define , respectively , @xmath150 and the @xmath85-dependence of @xmath157 in a spatially infinitesimal manner . now , we show that trps on the lapse function @xmath12 does not hold due to its inconsistency with the theta particle part of @xmath158 : trps is preserved for @xmath159 , which is equivalent to the trp invariance of @xmath147 , but non - zero @xmath160 breaks trps spontaneously . the crucial point here is that , due to the _ superposition form _ of the distribution @xmath158 and the singleness of its non - zero fine - grained level , for every theta or sen spin , it is in principle impossible to specify to which component @xmath157 this spin belongs , in the phase space region of overlap between @xmath122 and @xmath123 . this is the novel feature of the double lynden - bell distribution as an equilibrium distribution ( refer to figs.1 and 2)@xcite and is the mainstay of this paper . due to this point , in the ground state @xmath158 , trp , that is , a replacement of the clock in the theta particle system@xmath161is required to be applied for both proper times @xmath149 for @xmath126 by a single correspondence rule ( i.e. , a single time lapse ) between clock and proper time due to the singleness of @xmath158 . thus , one of the two proper times becomes an obvious constraint on trp for the other proper time , which breaks trps on the lapse function @xmath12 spontaneously in a collisionless equilibrium theta particle system . by the above step , once trps breaks spontaneously , the lapse function decomposes into @xmath163 for ground state expectation value @xmath164 and fluctuation @xmath165 . thus , the proper time @xmath135 also decomposes into @xmath166 for ground state expectation value @xmath167 and fluctuation @xmath168 . this decomposition in eq.([eq : tt ] ) arises because , according to the adm spatial coordinate transformation of the space - time @xmath169 with the signature @xmath170 , we set the proper time of the local theta particle system , which is , in the present case , also that of the matter system @xmath80 , as @xmath171 by employing an _ a priori _ and reversibly flowing time @xmath98 that is used in the adm decomposition just as a parameter . consequently , the trials of time increment @xmath172 fluctuate around the trial mean with a non - zero relative variance @xmath2 that sets the time scale of symmetry breaking . here , it is consistent to assume that the behavior of the time increment fluctuation is governed by a certain statistical law , which is independent from the model . in particular , its relative variance @xmath2 is to be treated as a physical constant of nature . as will be shown in the next part , in our model , the time @xmath173 within whose interval the state reduction of the matter system @xmath80 itself occurs is estimated by @xmath174 for the unit time scale @xmath175 of quantum mechanical phenomena in the matter system @xmath80 . so , to determine the value of @xmath2 , we require experimental comparisons with this formula for particular matter systems . in time - dependent processes of the matter system @xmath80 with its hamiltonian @xmath177 , the statistical variance of time increment trials is represented by the non - unitary time evolution of the state vector of the system @xmath80 in the schr@xmath178dinger picture , as seen in the following way . in the schr@xmath178dinger picture , the time - evolution equation of the state vector @xmath179 of the matter system @xmath80 is the schr@xmath178dinger equation . we write it in the time increment formalism using the proper time of both the local theta particle system and the matter system @xmath80 @xmath180 the average of its formal solution over time increment trials is @xmath181 when time increment trials @xmath172 are simply assumed to follow a normal distribution @xmath182 with mean @xmath124 and variance @xmath183 . this formula gives the time evolution of the state vector in the history representation : @xmath184 the first exponential factor gives unitary time evolution . to understand the meaning of the second exponential factor in a formal way , we note the following points . in quantum mechanics , a hamiltonian @xmath185 is an hermitian operator . so , for its real eigenvalues @xmath186 , there exists a unique spectral family @xmath187 . here , the spectral component @xmath188 is defined by the restriction of @xmath177 to the part which lies within the eigenspace for eigenvalue @xmath85 . the spectral decomposition of the hamiltonian is @xmath189 from the projection property of the spectral components @xmath190 it follows that @xmath191 due to this , on average , the second exponential factor in eq.([eq : sol ] ) forms a family with the properties of a contraction semi - group with parameter @xmath135 . now , it is clear that this factor causes the irreversible non - unitary time evolution of @xmath192 for energy eigenstates @xmath193 and complex coefficients @xmath194 . this time evolution is a quantum state reduction with respect to energy . indeed , when we describe this time evolution statistically , by solving the corresponding von neumann equation , it follows that the non - diagonal elements of the density matrix of an energy pure state are damped exponentially with time constant @xmath173 . finally , we extend this discussion to the case of the complex quantum system of the matter system @xmath80 and the sen spin system @xmath21 , or equivalently , @xmath195 , in the coherence domain of @xmath80 . in the quantum regime of the sen spin system , we need to introduce its state vectors , since orbit - less sen spinor fields @xmath21 and @xmath196 are replaced by quantum field operators of unconstrained schwinger bosons . as a remarkable point , the hilbert space of state vectors of the complex system @xmath197 , @xmath198 , decomposes into a direct integral of mutually _ unitarily inequivalent _ coherent subspaces ( i.e. , superselection sectors ) , @xmath199 , as @xmath200 where @xmath201 is the set of all ground state expectation values of the lapse function @xmath12 and the shift vector @xmath13 in the total spin system ( refer to footnote 48 ) and @xmath124 is the space - time diffeomorphism invariant functional measure . this unitary inequivalence is attributed to the spontaneous breaking of space - time diffeomorphism invariance in the total spin system where sen spin system has an infinite number of field degrees of freedom and infinitely many unitarily inequivalent coherent states exist where each state is specified by a pair of classical fields @xmath195@xcite . now , based on this fact , we assume an arbitrarily given observable @xmath202 , depended on by gravity , of a given quantum system @xmath203 with set of eigenstates @xmath204 and consider the complex system of @xmath203 and @xmath197 with the von neumann form interaction that is a result of gravity and the equivalence principle takes the form @xmath205 where @xmath206 acts only in the hilbert space of the system @xmath197 such that @xmath207 , the times evolved by @xmath206 , belong to different superselection sectors for each @xmath95@xcite . ] . then , state vectors of the total system @xmath208 in the von neumann superposition form , such as @xmath209 where @xmath210 and @xmath211 are distinct and no condition is imposed on @xmath212 or @xmath213 , automatically become mixtures , due to the superselection property of @xmath207 , within their own time scale such that , by it , the proper time difference reaches the planck time . moreover , if @xmath212 and @xmath213 are energy eigenstates with different eigenvalues , a measurement of @xmath202 of the system @xmath203 by the complex system @xmath197 follows from eq.([eq : sol ] ) as a state reduction in the matter system @xmath80 within a time scale @xmath173 . with this final step , our mechanism to derive a quantum mechanical non - unitary time - dependent process @xmath90 in the world for a macroscopically spatially coherent matter system @xmath80 via trps breaking @xmath214&(a)\ar[r]&(b)\ar[r]&(c)\ar[r]&(d)}\label{eq : summary}\ ] ] is complete . in our model , we have proposed to replace the scheme of the copenhagen school quantum mechanics ( q.m . ) of a self - contained matter system @xmath80 , that consists of unitary time - dependent process and non - unitary events , in the history description with time @xmath135@xcite @xmath215\|{(d^\prime)}}\ \ { \rm{in}}\ \ { \rm{q.m.}}\label{eq : qm}\ ] ] by the scheme shown in eq.([eq : int ] ) consisting of @xmath84 , @xmath85 and @xmath80 ( or even without @xmath80 , as noted at the end of this paper ) in the time increment description . ( in our interpretation of quantum mechanics , while the positions and momenta of the localizable second - quantized particles in @xmath80 are originally definite variables , their uncertainties are owned by space - like hypersurfaces at a time , via the momentum constraint in the case of momenta . these hypersurfaces unite with their background @xmath80 to form quantum eigenstates of the observable position or momentum under superposition . the observable position and momentum are mutually exclusively definite canonical variables . because the spin is independent from the spatial motion , we treat it separately . ) summarizing the contents of the diagram eq.([eq : int ] ) briefly , the local theta particle system @xmath84 , whose superposed form collisionless equilibrium ( i.e. , double lynden - bell ) distribution is dynamically generated from the initial distribution prepared by the matter system @xmath80 , provided it is coherent , via the coupling of its vector part with the shift vector , converts the reversibly flowing time @xmath98 into an irreversibly flowing ( see the argument below ) proper time @xmath135 with increments fluctuating within the world for the matter system @xmath80 by the spontaneous breaking of trps due to the existence of the trp invariant double lynden - bell potential eq.([eq : pot ] ) for the time lapse ( spatial diffeomorphism invariance is also broken spontaneously ) ; the time lapse is rescaled in the ground state eq.([eq : gstate ] ) in two ways . by solving the schr@xmath178dinger or the von neumann equation written in the time increment formalism ( e.g. , eq.([eq : sch ] ) ) , we have shown that this conversion of time results in non - unitary events , which include gravitational state mixings and measurements , in the world for the matter system @xmath80 , with irreversibility arising from using the time fluctuations @xmath168 in eq.([eq : tt ] ) but expressed by using the _ a priori _ reversible temporal flow @xmath98 . linking to the timeless property of canonical quantum gravity@xcite , it is to be noted that , in the trps unbroken phase , quantum mechanical unitary processes as temporal steps with a time measure are in principle indistinguishable due to the gauge symmetry nature of trps . when a local theta particle system induces the trps broken phase , there arises a definite time measure due to @xmath167 in eq.([eq : tt ] ) and these processes become distinguishable . it is worthwhile to briefly compare our proposal for resolving the measurement problem with a few other proposals that have formal completeness in the description by a state vector . first , the essential difference from the program of quantum decoherence@xcite is that , while that approach must describe the process statistically using the mixed state of the quantum system with a density matrix , our mechanism works fundamentally with pure states . second , compared with the continuous spontaneous collapse model in ref.@xcite , the advantage of our model is that , while the evolution law of the continuous spontaneous collapse model is modified _ a priori _ , ours is not except for the scale of symmetry breaking ; in particular , our time - evolution equation is not modified from the schr@xmath216dinger equation . third , the novel result that definitively distinguishes our proposal from others is a consistent introduction of the universal scheme of quantum measurements shown in eq.([eq : int ] ) as an extension of canonical gravity theory . finally , a particularly novel feature of our proposal is that , in the mechanism invoking gravity , the main source of its properties is not gravity itself but a particle , that is , the theta particle , which we hypothesize to be cold dark matter . this feature and our gravitational state mixing mechanism distinguish our proposal from the di@xmath217si - penrose proposal _ for quantum state reduction_@xcite . first , we have argued that a matter system @xmath80 may produce a marked double lynden - bell distribution of theta particles . however , in reverse , a marked double lynden - bell distribution of theta particles does _ not always _ require a matter system that produces it . indeed , the higher chemical potential part of the marked double lynden - bell distribution of a local theta particle system , comparable to the cyan domains in figs . 1 and 2 , covers a highly improbable region of the phase space from the aspect of lynden - bell statistics , comparable to the gray curves in figs . 1 and 2 , and once it is generated , this part is trapped in this phase space region until it evaporates@xcite and is unable to thermalize with the rest of the system by the regular local theta particle dynamics in eq.([eq : xy])@xcite . thus , the local theta particle system still induces the broken phase of space - time diffeomorphism invariance , even after the matter system @xmath80 is lost . this observation shows that , to be accurate , _ the world for a matter system @xmath80 _ used so far must be generalized to _ the world for a local theta particle system @xmath84_. second , in the quantum mechanical regime of theta particle system , there is pairing instability in the ground state , since the theta particle is a fermion and its self - interactions through the aligned spin are attractive@xcite . in such a situation , we expect the realization of superconductivity - like macroscopically coherent behavior of quantum theta particles evading the pauli exclusion principle , below a certain critical temperature . if our proposed mechanism can be tested experimentally , investigating this aspect of our model , especially determining the scale of the critical temperature at the same time as we primarily determine the scales of coupling strengths and mass of theta particles and the scale of trps breaking , will be a significant issue . 99 m. born , zeitschrift . f@xmath218r . physik . * 38 * , 803 ( 1926 ) ; n. bohr , nature . * 121 * , 580 , ( 1928 ) . j. von neumann , math . 104 * , 570 ( 1931 ) ; _ mathematische grundlagen der quantenmechanik _ ( springer - verlag , berlin , 1932 ) ; e. p. wigner , am . j. phys . * 31 * , 6 ( 1963 ) . l. de broglie , j. phys . radium . * 8 * , 225 ( 1927 ) . d. bohm , phys . 85 * , 166 ( 1952 ) ; _ ibid_. * 85 * , 180 , ( 1952 ) . h. everett , rev . * 29 * , 454 ( 1957 ) ; j. a. wheeler , _ ibid_. * 29 * , 463 ( 1957 ) . h. d. zeh , found . * 1 * , 69 ( 1970 ) . w. h. zurek , phys . d * 24 * , 1516 ( 1981 ) ; _ ibid_. * 26 * , 1862 ( 1982 ) ; rev . phys . * 75 * , 715 ( 2003 ) . e. joos and h. d. zeh , z. phys . b. * 59 * , 223 ( 1985 ) . g. c. ghirardi , a. rimini and t. weber , phys . d * 34 * , 470 ( 1986 ) ; e. joos , _ ibid_. * 36 * , 3285 ( 1987 ) ; g. c. ghirardi , a. rimini and t. weber , _ ibid_. * 36 * , 3287 ( 1987 ) . a. bassi and g. c. ghirardi , phys . rep . * 379 * , 257 ( 2003 ) . f. k@xmath219rolyh@xmath219zy , nuovo cimento * 52 * , 390 ( 1966 ) . l. di@xmath217si , phys . a * 120 * , 377 ( 1987 ) ; phys . a * 40 * , 1165 ( 1989 ) . r. penrose , gen . * 28 * , 581 ( 1996 ) . a. sen , j. math . * 22 * , 1781 ( 1981 ) . a. sen , int . * 21 * , 1 ( 1982 ) . a. sen , phys . b * 119 * , 89 ( 1982 ) . j. a. wheeler , in _ battelle rencontres _ , c. de witt and j. a. wheeler ( eds . ) ( benjamin , new york , 1968 ) . b. s. de witt , phys . rev . * 160 * , 1113 ( 1967 ) . a. vilenkin , phys . d * 39 * , 1116 ( 1989 ) . w. g. unruh and r. m. wald , phys . d * 40 * , 2598 ( 1989 ) . c. j. isham , in _ integrable systems , quantum groups , and quantum field theories _ , l. a. ibort and m. a. rodriguez ( eds . ) ( kluwer , london , 1993 ) . a. m. dirac , proc . a * 246 * , 326 ( 1958 ) ; _ ibid_. * 246 * , 333 ( 1958 ) ; phys . rev . * 114 * , 924 ( 1959 ) . r. arnowitt , s. deser and c. w. misner , in _ gravitation : an introduction to current research _ , l. witten ( ed . ) ( wiley , new york , 1962 ) . k. kucha@xmath220 , j. math . * 17 * , 777 ( 1976 ) . c. j. isham and k. kucha@xmath220 , ann . phys . * 164 * , 288 ( 1985 ) . j. m. nester , phys . a * 83 * , 241 ( 1981 ) . r. geroch , j. math . phys . * 13 * , 956 ( 1972 ) . m. bojowald , _ canonical gravity and applications : cosmology , black holes , and quantum gravity _ ( cambridge university press , cambridge , 2010 ) . t. d. newton and e. p. wigner , rev . mod . * 21 * , 400 ( 1949 ) ; a. s. wightman , _ ibid_. * 34 * , 845 ( 1962 ) . coherent states of photons can be localized without any limitations . for these states , the full localization of the photon number density can be obtained . see i. bialynicki - birula and z. bialynicka - birula , phys . a * 79 * , 032112 ( 2009 ) for details . l. mandelstam and i. g. tamm , j. phys . ( ussr ) * 9 * , 249 ( 1945 ) . a. messiah , _ quantum mechanics . ( north - holland , amsterdam , 1972 ) . d. lynden - bell , mon . not . . soc . * 136 * , 101 ( 1967 ) . e. konishi and m. sakagami , phys . e * 91 * , 032144 ( 2015 ) ; for a comprehensive review , see e. konishi , int . b * 30 * , 1630007 ( 2016 ) . y. levin , r. pakter and f. b. rizzato , phys . e * 78 * , 021130 ( 2008 ) . r. pakter and y. levin , phys . lett . * 106 * , 200603 ( 2011 ) . y. levin , r. pakter , f. b. rizzato , t. n. teles and f. p. c. benetti , phys . rep . * 535 * , 1 ( 2014 ) . a. campa , t. dauxois , d. fanelli and s. ruffo , _ physics of long - range interacting systems _ ( oxford university press , oxford , 2014 ) . a. campa , t. dauxois and s. ruffo , phys . rep . * 480 * , 57 ( 2009 ) . w. braun and k. hepp , comm . * 56 * , 101 ( 1977 ) . for details , see chapter 3 of h. umezawa , _ advanced field theory : micro , macro and thermal physics _ ( american institute of physics , new york , 1993 ) . e. joos , in _ decoherence and the appearance of a classical world in quantum theory _ , d. giulini _ et al . _ ( eds . ) ( springer - verlag , berlin , 1996 ) . r. bachelard , c. chandre , d. fanelli , x. leoncini and s. ruffo , phys . * 101 * , 260603 ( 2008 ) . f. p. c. benetti , t. n. teles , r. pakter and y. levin , phys . lett . * 108 * , 140601 ( 2012 ) . l. n. cooper , phys . rev . * 104 * , 1189 ( 1956 ) .	we propose a mechanism for time reparametrization symmetry breaking in canonical gravity . we consider a model of spinor gravity , based on sen s reformulation of canonical gravity as a spin system , with one additional long - range self - interacting massive spinor particle that is coupled to spinor gravity . the symmetry breaking is identified with the origin of the quantum mechanical non - unitary evolution . a part of our approach to deriving non - unitary time - dependent processes of state vectors uses an idea from the penrose argument for gravity s position in quantum mechanics .
here we present in detail the procedure that we use to determine the minimum curvature of the free energy functional for a given temperature . for simplicity we concentrate on the bethe lattice . the ginzburg - landau free energy functional @xmath70 in the hilbert space of the matsubara green s functions @xmath71 is given by @xcite @xmath72 where the first term is the energy cost of forming the weiss field @xmath73 around a given site , while the second term is the free energy of an electron at this site in the presence of the weiss field . close to a local extremum @xmath74 , we can taylor expand @xmath70 in terms of deviation from this point @xmath75 : @xmath76 where @xmath77 we introduce a gradient function @xmath78 and define an iteration - substitution procedure by @xmath79 which gives the minimum of the free energy as the iteration step @xmath80 . in the case of the free energy functional given by eq . ( [ eq1 ] ) , we have @xmath81 and in the iterative procedure the green s function converges to the minimum of the free energy which is at the same time also a self - consistent solution of the dmft equations , given by the relation @xmath82 . using ipt impurity solver , panel ( a ) , and ctqmc impurity solver , panel ( b ) . the dashed lines in panel ( b ) are linear fits to the data . the insets are the corresponding eigenvalues determined by the slopes from the main panels.,width=307 ] the curvature of the free energy for interaction @xmath13 and temperature @xmath5 can be obtained as follows . the eigenbasis @xmath83 and eigenvalues @xmath84 of matrix @xmath85 are defined by @xmath86 we can expand @xmath87 as @xmath88 where @xmath89 are the coefficients of @xmath87 in the eigenvalue basis . substituting into eq . ( [ eq5 ] ) , one obtains @xmath90 where @xmath91 , @xmath92 and @xmath93 , @xmath94 ) . red dashed line is the instability line @xmath95 , full red line is the line of the first order mit , and green and blue dotted lines are left and right spinodals . , width=403 ] for large @xmath33 the term with lowest @xmath96 , which corresponds to the lowest eigenvalue @xmath97 , is dominant @xmath98 here @xmath99 is the coefficient corresponding to the green function with the lowest eigenvalue @xmath31 . now it is obvious that through iterations , the solution @xmath100 approaches to @xmath101 exponentially along a direction defined by the eigenvector of @xmath85 corresponding to its minimal eigenvalue @xmath31 . the coefficient @xmath102 and the corresponding eigenvalue @xmath31 are then obtained from the slope in the iterative relation @xmath103 = const - nb_{\alpha_0},\ ] ] which follows from eq . ( [ eq9 ] ) . in practice , to obtain @xmath31 ( and thus the curvature of free energy ) , we monitor dmft loop convergence rate , @xmath104 , in as many iterations as possible and then linearly fit @xmath105 versus iteration index @xmath33 . here @xmath106 is the lowest matsubara frequency . for small @xmath31 , @xmath107 . we repeat this procedure for different values of @xmath13 at the same temperature @xmath5 to determine @xmath95 in which @xmath108 is minimal . it takes few iterations of the dmft loop to enter into the linear regime given by eq . ( [ eq12 ] ) . with ipt impurity solver , we can use data from several tens of iterations to determine the slope @xmath102 , fig . the solution with ctqmc impurity solver has a statistical error and the number of iterations is limited before the difference @xmath109 becomes too small and acquires a large relative statistical error . nevertheless , we were able to determine rather precisely the eigenvalue @xmath31 and the interaction @xmath95 for which it becomes minimal , fig . the `` instability line '' corresponding to the minimum curvature of the free energy is shown in fig . 2(a ) ( ipt phase diagram ) , and fig . 2(b ) ( ctqmc phase diagram ) . error bars in fig . 2(b ) are estimates of the uncertainty in the position of the instability line . the resistivity @xmath110 is calculated along the lines parallel to the instability line @xmath95 . here the resistivity is fist divided by its value @xmath112 at @xmath113 , fig . then for each @xmath114 the temperature axis is scaled by @xmath115 where the scaling parameter @xmath115 is chosen to collapse the data onto two branches : insulating - like for @xmath43 and metallic - like for @xmath116 . the scaling was done in such a way that data were collapsed on the lowest curves with @xmath117 as shown in fig . the scaling parameter @xmath115 has a power law form @xmath47 , where the prefactor @xmath118 depends on this referent value of @xmath13 . our data exhibit a reflection symmetry which is seen in fig . 4(a ) , where the resistivity @xmath63 ( for @xmath43 ) and conductivity @xmath64 ( @xmath45 ) can be mapped onto each other by reflection with @xmath120 . vs. @xmath114 obtained from scaling procedure shown on fig . ( b ) @xmath52 vs. @xmath114 after rescaling by @xmath122 . the boundary of the quantum critical region is now given by condition @xmath66 . , width=307 ] the standard estimate for the scale ( prefactor ) of the crossover temperature is obtained by requiring that the scaling variable @xmath123 at the point where the scaling function changes its functional form ; in our case this corresponds to the temperature below which the mirror symmetry of the scaling curves no longer holds . before rescaling the prefactor , this is found at @xmath124 . our final form of scaled data is shown in fig . 4(b ) , where @xmath121 - axis is rescaled by @xmath125 so that @xmath66 sets the _ boundary of the quantum critical region_. the scaling parameter @xmath52 as function of @xmath114 is shown in fig . 5(a ) ( before rescaling with @xmath126 ) and in fig . 5(b ) ( after rescaling with @xmath126 ) . the corresponding values for @xmath118 and @xmath127 from the power law fit are also given . we can now plot the crossover temperature @xmath52 setting the boundary of qc region on our phase diagram . @xmath128 condition is equivalent to @xmath129 where @xmath130 is value of @xmath13 at temperature @xmath131 , along the instability line . this equation implicitly defines the crossover line @xmath132 . alternatively , we can invert this dependence to describe the same crossover line as @xmath133 that takes the form @xmath134 as we can see from this expression , the crossover line approaches the instability line at low @xmath5 and diverges away from it at high @xmath135 the phase diagram including the instability line and the crossover temperature @xmath115 is shown on fig . 6 .	we perform a systematic study of incoherent transport in the high temperature crossover region of the half - filled one - band hubbard model . we demonstrate that the family of resistivity curves displays characteristic quantum critical scaling of the form @xmath0 , with @xmath1 , and @xmath2 . the corresponding @xmath3-function displays a `` strong coupling '' form @xmath4 , reflecting the peculiar mirror symmetry of the scaling curves . this behavior , which is surprisingly similar to some experimental findings , indicates that mott quantum criticality may be acting as the fundamental mechanism behind the unusual transport phenomena in many systems near the metal - insulator transition . many systems close to the metal - insulator transition ( mit ) often display surprisingly similar transport features in the high temperature regime @xcite . here , the family of resistivity curves typically assumes a characteristic `` fan - shaped '' form ( see fig . 1(a ) ) , reflecting a gradual crossover from metallic to insulating transport . at the highest temperatures the resistivity depends only weakly on the control parameter ( concentration of charge carriers @xcite , or pressure @xcite ) , while as @xmath5 is lowered , the system seems to `` make up its mind '' and rapidly converges towards either a metallic or an insulating state . since temperature acts as a natural cutoff scale for the metal - insulator transition , such behavior is precisely what one expects for quantum criticality @xcite . in some cases @xcite , the entire family of curves displays beautiful scaling behavior , with a remarkable `` mirror symmetry '' of the relevant scaling functions @xcite . but under which microscopic conditions should one expect such scaling phenomenology ? what is the corresponding driving force for the transitions ? despite recent progress , such basic physics questions remain the subject of much ongoing controversy and debate . the phenomenon of disordered - driven anderson localization of noninteracting electrons is at present rather well understood based on the scaling formulation @xcite , and is generally viewed as an example of a @xmath6 quantum phase transition ( qpt ) . on the other hand , a considerable number of recent experiments @xcite provide compelling evidence that strong correlation effects - some form of mott localization - may be the dominant mechanism @xcite . should one expect similar or very different transport phenomenology in the mott picture ? is the paradigm of quantum criticality even a useful language to describe high temperature transport around the mott point ? these issues are notoriously difficult to address , because conventional fermi liquid concepts simply can not be utilized in the relevant high temperature _ incoherent _ regime . in this letter we answer this question in the framework of dynamical mean - field theory ( dmft ) @xcite , the only theoretical method that is most reliable precisely at high temperatures . with respect to the instability line @xmath7 ( black dashed line , see the text ) . data are obtained using ipt impurity solver . ( b ) resistivity scaling ; essentially identical scaling functions are found from ctqmc ( open symbols ) and from ipt ( closed symbols),width=316 ] _ model _ _ and dmft solution_. we consider a single - band hubbard model at half - filling @xmath8 where @xmath9 and @xmath10 are the electron creation and annihilation operators , @xmath11 , @xmath12 is the hopping amplitude , and @xmath13 is the repulsion between two electrons on the same site . we use a semicircular density of states , and the corresponding half - bandwidth @xmath14 is set to be our energy unit . we focus on the paramagnetic dmft solution , which is formally exact in the limit of large coordination . here the hubbard model maps onto an effective anderson impurity model supplemented by a self - consistency condition @xcite . to solve the dmft equations we use the iterated perturbation theory ( ipt ) @xcite and cross - check our results with numerically exact continuous time quantum monte carlo ( ctqmc ) @xcite . we find , in agreement with previous work @xcite , that after appropriate energy rescaling ( see below ) , the two methods produce qualitatively and even quantitatively identical results in the incoherent crossover region that we examine . and insulating @xmath15 spinodals ( dotted lines ) are found at @xmath16 ; the corresponding first order phase transition is shown by a thick solid line . the thick dashed line , which extends at @xmath17 shows the `` instability trajectory '' @xmath18 , and the crossover temperature @xmath19 delimits the qc region ( dash - dotted lines ) . the inset shows examples of eigenvalue curves at three different temperatures , with pronounced minima at @xmath18 determining the instability trajectory . , width=326 ] it is well known that at very low temperatures @xmath20 , this model features a first order metal - insulator transition terminating at the critical end - point @xmath21 ( fig . 2 ) , very similar to the familiar liquid - gas transition @xcite . for @xmath17 , however , different crossover regimes have been tentatively identified @xcite but they have not been studied in any appreciable detail . the fact that the first order coexistence region is restricted to such very low temperatures provides strong motivation to examine the high temperature crossover region from the perspective of `` hidden quantum criticality '' . in other words , the presence of a coexistence dome at @xmath22 , an effect with very small energy scale , is not likely to influence the behavior at much higher temperatures @xmath23 . in this high temperature regime smooth crossover is found , which may display behavior consistent with the presence of a `` hidden '' quantum critical ( qc ) point at @xmath6 . to test this idea , we utilize _ standard _ scaling methods appropriate for quantum criticality , and compute the resistivity curves along judiciously chosen trajectories respecting the symmetries of the problem . _ instability trajectory formalism_. previous work has already recognized @xcite that , in order to reveal the proper scaling behavior close to the critical end - point , one has to follow a set of trajectories parallel to `` zero field '' trajectory @xmath24 . we thus expect @xmath25 to play the role of the scaling variable corresponding to a symmetry - breaking field favoring one of the two competing ( metal vs. insulator ) phases . by analogy @xcite to the familiar liquid - gas transition , we determine the precise location of such an `` instability trajectory '' by examining the curvature of the corresponding free energy functional @xcite . this curvature vanishes at @xmath21 and is finite and _ minimal _ at @xmath26 along this instability line . consequently , as in refs . @xcite , our problem is recast as an eigenvalue analysis of the corresponding free energy functional @xmath27 $ ] for which the dmft green s function solution @xmath28 represents a local extremum , and can be regarded as a vector in an appropriate hilbert space . the free energy near such an extremum can be written as @xmath27=\mathcal{f}_{0}+tt^{2}\sum_{m , n}\delta g(i\omega_{m})m_{mn}\delta g(i\omega_{n})+\dots$ ] , where @xmath29}{\partial g(i\omega_{m})\partial g(i\omega_{n})}\right\|_{g = g_{dmft}}\ ] ] and @xmath30 . the curvature of the free energy functional is determined by the lowest eigenvalue @xmath31 of the _ fluctuation matrix _ @xmath32 . as explained in supplementary notes @xcite , @xmath31 can be obtained from the iterative solution of dmft equations . the difference of the green s functions in iterations @xmath33 and @xmath34 of the dmft self - consistency loop is given by @xmath35 and therefore @xmath31 determines the rate of convergence of the green function to its solution . an example of our calculations is shown in the inset of fig . 2 , where the eigenvalues at several temperatures are plotted as a function of interaction @xmath36 . the minima of these curves define the locus of the `` _ _ instability trajectory _ _ '' @xmath18 , which terminates at the critical end - point @xmath37 , as shown on fig . 2 . note , that the immediate vicinity @xmath38 of the critical end - point has been carefully studied theoretically @xcite , and even observed in experiments @xcite , revealing classical ising scaling ( since one has a finite temperature critical point ) of transport in this regime . in our study , we examine the crossover behavior at much higher temperatures @xmath23 , displaying very different behavior : precisely what is expected in presence of _ quantum criticality_. _ resistivity calculation . _ to reveal quantum critical scaling , we calculate the temperature dependence of the resistivity along a set of trajectories parallel to our instability trajectory ( fixed @xmath39 ) . resistivity was calculated using standard dmft procedures @xcite , with the maximum entropy method @xcite utilized to analytically continue ctqmc data to the real axis . the resistivity results are shown in fig . 1 , where on panel ( a ) ipt resistivity data for @xmath40 in the temperature range of @xmath41 are presented ( ctqmc data are not shown for the sake of clarity of the figure ) . the resistivity is given in units of @xmath42 , maximal resistivity according to the boltzmann quasi - classical theory of transport @xcite . the family of resistivity curves above ( @xmath43 ) the `` separatrix '' @xmath44 ( dashed line , corresponding to @xmath7 ) has an insulating - like behavior , while metallic dependence is obtained for @xmath45 . ) as function of control parameter @xmath46 ; the inset illustrates power - law dependence of scaling parameter @xmath47 ; ( b ) resistivity @xmath44 of the `` separatrix '' . excellent agreement is found between ipt ( closed symbols ) and ctqmc ( open symbols ) results . , width=316 ] _ scaling analysis_. according to what is generally expected for quantum criticality , our family of curves should satisfy the following scaling relation:@xmath48 we thus first divide each resistivity curve by the `` separatrix '' @xmath49 and then rescale the temperature , for each curve , with an appropriately chosen parameter @xmath50 to collapse our data onto two branches , fig . 1(b ) . note that this unbiased analysis _ does not assume _ any specific form of @xmath51 : it is determined for each curve simply to obtain optimum collapse of the data @xcite . this puts us in a position to perform a stringent test of our scaling hypothesis : true quantum criticality corresponds to @xmath51 which vanishes at @xmath7 and displays power law scaling with the same exponents for both scaling branches . as seen in fig . 3(a ) @xmath52 falls sharply as @xmath53 is approached , consistent with the qc scenario , but opposite to what is expected in a `` classical '' phase transition . the inset of fig . 3(a ) with log - log scale shows clearly a power - law behavior of @xmath47 , with the estimated power @xmath54 for `` metallic '' side , and @xmath55 for an insulating branch . we also find ( fig . 3(b ) ) a very unusual form of our critical resistivity @xmath44 , corresponding to the instability trajectory . its values largely exceeds the mott limit , yet it displays metallic - like but non - fermi liquid - like temperature dependence @xmath56 such puzzling behavior , while inconsistent with any conventional transport mechanism , has been observed in several strongly correlated materials close to the mott transition @xcite . our results thus suggest that it represents a generic feature of mott quantum criticality . _ @xmath57function and mirror symmetry of scaled curves . _ to specify the scaling behavior even more precisely , we compute the corresponding @xmath3-function @xcite @xmath58 , with @xmath59 being the inverse resistivity scaling function . remarkably ( fig . 4(a ) ) , it displays a nearly linear dependence on @xmath60 , and is continuous through @xmath7 indicating precisely the same form of the scaling function on both sides of the transition - another feature exactly of the form expected for genuine quantum criticality . this functional form is very natural for the insulating transport , as it is obtained even for simple activated behavior @xmath61 @xmath62 . the fact that the same functional form _ persists _ well into the metallic side is a surprise , especially since it covers almost an order of magnitude for the resistivity ratio . such a behavior has been interpreted @xcite to reflect the strong coupling nature of the critical point , which presumably is governed by the same physical processes that dominate the insulator . this points to the fact that our qc behavior has a strong coupling , i.e. non - perturbative character . the fact that the @xmath57function assumes this logarithmic form on both sides of the transition is mathematically equivalent @xcite to stating that the two branches of the corresponding scaling functions display mirror symmetry over the same resistivity range . indeed , we find that transport in this qc region exhibits a surprisingly developed reflection symmetry ( dash vertical lines of fig . 4(a ) mark its boundaries ) . such a symmetry is clearly seen in fig . 4(b ) , where the resistivity @xmath63 ( for @xmath43 ) and conductivity @xmath64 ( @xmath45 ) can be mapped onto each other by reflection with @xmath65 @xcite . note that @xmath66 sets the boundary of the quantum critical region , over which the reflection symmetry of scaled curves is observed . it is depicted by dash - dotted crossover lines @xmath52 on phase diagram of fig . 2 @xcite . these remarkable features of the @xmath57function , and associated reflection - symmetry , have been observed earlier in experimental @xcite and theoretical @xcite studies , which tentatively associated this with disorder - dominated mits . speculation that @xmath67 reveals disorder as the fundamental driving force for mit , presumably reflects the fact that , historically , it has first been recognized for anderson transitions @xcite . our work , however , shows that such behavior can be found even in absence of disorder - in interaction driven mits . this finding calls for re - thinking of basic physical processes that can drive the mit . function shows linear in @xmath68 behavior close to the transition . open symbols are for metallic branch ( @xmath45 ) and closed ones are for the insulating side ( @xmath43 ) ; vertical dashed lines indicate the region where mirror symmetry of curves is found . ( b ) reflection symmetry of scaled curves close to the transition.,width=316 ] _ conclusions : _ we have presented a careful and detailed study of incoherent transport in the high temperature crossover regime above the critical endpoint @xmath21 of a single - band hubbard model . our analysis revealed a so - far overlooked scaling behavior of the resistivity curves , which we interpreted as evidence of `` hidden '' mott quantum criticality . precisely locating the proposed qcp in our model is hindered by presence of the low temperature coexistence dome , which limits our quantum critical scaling to the region well above @xmath69 regarding the nature of transport in the qc regime , we found that the critical resistivity well exceeds the mott limit , and yet it - surprisingly - assumes a metallic form , in dramatic contrast to conventional mit scenarios . these features , together with large amounts of entropy characterizing this entire regime @xcite , prove surprisingly reminiscent of the holographic duality scenario @xcite for a yet - unspecified qc point . interestingly , the holographic duality picture has - so far - been discussed mostly in the context of quantum criticality in correlated metals ( e.g. @xmath6 magnetic transitions in heavy fermion compounds ) . ours is the first work proposing that the same physical picture could apply to quantum criticality found at the mit . we believe that our results provide a significant new perspective on qc around the mott transition , and a deeper understanding of an apparent universality in the high temperature crossover regime . our method traces a clear avenue for further searches for qc scaling , which are likely to be found in many other regimes and models . in particular , it would be interesting to study a corresponding critical regime by going beyond the single - site dmft analysis . it was shown in @xcite that inclusion of spatial fluctuations does not significantly modify the high temperature crossover region in half - filled hubbard model . consequently , we expect our main findings to persist . an even more stringent test of our ideas should be provided in models where the critical end - point @xmath21 can be significantly reduced . this may include studies of the mott transition away from half - fillling @xcite , or in systems with frustrations @xcite . in such situations the proposed scaling regime should extend to much lower temperatures , perhaps revealing more direct evidence of the - so far - hidden mott qc point . our ideas should also be tested by performing more detailed transport experiments in the relevant incoherent regime , a task that may be easily accessible in various organic mott systems @xcite , where @xmath21 is sufficiently below room temperature . the authors thank k. haule for the usage of his ctqmc code , and m. jarrell for the use of his maximum entropy code for analytical continuation of the ctqmc data . this work was supported by the national high magnetic field laboratory and the nsf through grants dmr-0542026 and dmr-1005751 ( h.t . and v.d . ) , and serbian ministry of science under project no . on171017 ( j.v . and d.t . ) . d.t . acknowledges support from the nato science for peace and security programme grant no . eap.rig.983235 . numerical simulations were run on the aegis e - infrastructure , supported in part by fp7 projects egi - inspire , prace-1ip and hp - see . 26 e. abrahams _ et al . _ , rev . mod . phys . * 73 * , 251 ( 2001 ) . p. limelette _ et al . _ , science * 302 * , 89 ( 2003 ) . f. kagawa _ et al . _ , nature * 436 * , 534 ( 2005 ) . v. dobrosavljevi _ et al . _ , phys . rev . lett . * 79 * , 455 ( 1997 ) . e. abrahams _ et al . _ , phys . rev . lett . * 42 * , 673 ( 1979 ) . a. camjayi _ et al . , _ nature phys . * 4 * , 932 ( 2008 ) . a. georges _ et al . _ , rev . mod . phys . * 68 * , 13 ( 1996 ) . p. werner _ et al . _ , phys . rev . lett . * 97 * , 076405 ( 2006 ) . k. haule , phys . rev . b * 75 , * 155113 ( 2007 ) . g. kotliar , _ et al . _ , phys . rev . lett . * 48 * , 5180 ( 2000 ) . m. j. rozenberg _ et al . _ , phys . rev . lett . * 75 * , 105 ( 1995 ) . j. h. mooij , phys . status solidi , a * 17 * , 521 ( 1973 ) . g. moeller _ et al . _ , phys . rev . b * 59 * , 6846 ( 1999 ) . m. j. case and v. dobrosavljevi , phys . rev . lett . * 99 * , 147204 ( 2007 ) . see supplementary notes for eigenvalue analyis and scaling procedure . m. jarrell and j.e . gubernatis , physics reports * 269 * , 133 , ( 1996 ) . n. e. hussey _ et al . _ , philos . mag . 84 , 2847 ( 2004 ) . other criteria have also been emplyed to identify a mit crossover line . @xcite used the flatness of the lowest matsubara points of the self - energy , @xcite used the inflection point of the resistivity curves . we have not reveal any apparent scaling behavior following these trajectories . h. park _ et al . _ , phys . rev . lett . * 101 * , 186403 ( 2008 ) . b. j. powell and r. h. mckenzie , j. phys . : condens . matter 18 , r827 ( 2006 ) . d. simonian _ et al . _ , phys . rev . b. * 55 * , 13 421 ( 1997 ) . l. de leo _ et al . _ , arxiv:1009.2761v1 ( 2010 ) . s. sachdev , phys . rev . lett . * 105 * , 151602 ( 2010 ) . j. mcgreevy , physics * 3 * , 83 ( 2010 ) . g. sordi _ et al . _ , phys . rev . b * 80 * , 035129 ( 2009 ) . m. eckstein _ et al . _ , phys . rev . b. * 75 * , 125103 ( 2007 ) .
let @xmath2 be a compact riemannian manifold of dimension @xmath3 and let @xmath4 be the ( positive ) laplace operator on functions , which in local coordinates is given by @xmath5 here @xmath6 denotes the determinant of the metric tensor and @xmath7 are the components of the dual metric on the cotangent bundle . this operator is formally self - adjoint on the space of smooth functions with inner product @xmath8 defined by the riemannian measure @xmath9 . then @xmath4 extends to a self - adjoint operator @xmath10 with compact resolvent . this means there exists an orthonormal basis @xmath11 in @xmath12 consisting of eigenfunctions @xmath13 which we assume to be ordered such that @xmath14 . it is a classical problem to compute the eigenvalues and the eigenfunctions of the laplace operator on a given manifold . its solution allows complete control over the functional calculus and thus over solutions to the heat equation , the wave equation , or the schrdinger equation . in this paper we show how an adaptation of the method of particular solutions can be used to compute eigenvalues on manifolds with a high accuracy , beyond of what can currently be achieved by finite element methods or boundary element methods . we will focus primarily on the case of two dimensional oriented surfaces of constant curvature . these are the simplest examples of manifolds with non - trivial spectral geometry . they are topologically classified by their genus . by the gauss - bonnet theorem the curvature @xmath15 of a surface of genus @xmath16 can always be normalized to be @xmath17 by multiplying the metric by a positive constant . this divides oriented surfaces of constant curvature into three categories * @xmath18 : up to isometries there is only one such manifold : the round sphere @xmath19 . the spectrum of the laplace operator on the round sphere is of course well known and the spherical harmonics provide a basis consisting of eigenfunctions . * @xmath20 : such manifolds are isometric to flat tori which are obtained as quotients of @xmath21 by co - compact lattices . the moduli space of equivalence classes of flat metrics on a given topological torus is the modular surface @xmath22 , where @xmath23 is the upper half space . for a surface @xmath24 a basis consisting of eigenfunctions is obtained by taking the fourier modes associated with points in the dual lattice @xmath25 and the corresponding eigenvalues are the squares of the norms of these points . * @xmath26 : these are the so called hyperbolic surfaces . they can be obtained either as quotients @xmath27 by co - compact hyperbolic lattices in @xmath28 or by glueing hyperbolic pairs of pants . the moduli space of hyperbolic metrics on a given topological surface is a quotient of the teichmller space @xmath29 by a discrete group , the mapping class group . the teichmller space for a genus @xmath16 surface has dimension @xmath30 so that a constant curvature metric on a topological surface may be given by specifying @xmath30 parameters . unlike in the previous two cases the spectrum of the laplace operator of a hyperbolic surface can not currently be explicitly computed . the main purpose of this article is to establish an algorithm to compute the eigenvalues and eigenfunctions of the laplace operator on a hyperbolic surface up to a certain precision and with mathematically rigorous error bounds . once this is achieved for a number of eigenvalues it is then possible to compute values of the spectral zeta function and the spectral determinant of the laplace operator on such surfaces . we demonstrate how this can be done using the selberg trace formula in such a way that the error can again be bounded rigorously . the method we use is adapted to a pants decomposition of the surface . for a particular basis of functions we establish various bounds and estimates that prove that the method of particular solutions can be applied to an interval and yields all eigenvalues in that interval . in particular we prove a novel estimate that allows us to show non - existence of eigenvalues in a certain interval . in section [ mps ] we describe how an adaptation of the method of particular solutions can be applied to manifolds . the method uses a finite dimensional space of test functions that solve the eigenvalue equation with eigenvalue @xmath31 on open submanifolds that are glued together along co - dimension one hypersurfaces . the glueing conditions for the eigenfunctions , continuity of the eigenfunctions and their normal derivatives , is measured in terms of the differences of the function values and normal derivatives along the hypersurfaces . for a test function @xmath32 we introduce in equation ( [ functionaldef ] ) the number @xmath33 that measures these differences in suitable sobolev norms . we prove that if @xmath34 and @xmath33 is small then @xmath31 must be close to an eigenvalue . our theorem [ main01 ] together with the subsequent estimates on the constants gives a quantitative version of this statement for hyperbolic surfaces that are glued along geodesic segments . in section [ sec : surfaces ] we review the construction of hyperbolic surfaces from pairs of pants . this gives an explicit parametrization of teichmller space in terms of fenchel - nielsen coordinates and shows that hyperbolic surfaces can be cut open along geodesic segments into pairs of pants and subsequently into subsets of hyperbolic cylinders . in section [ sec : mps_on_surface ] we describe a set of basis functions for a surface of genus @xmath16 that is decomposed into pairs of pants . we prove that true eigenfunctions can be approximated by linear combinations of this set of basis functions with explicit bounds on the error . section [ algo ] deals with the construction of @xmath35 matrices @xmath36 , @xmath37 and @xmath38 with the following properties . 1 . the distance of @xmath31 to the spectrum can be bounded from above in terms of the first singular value @xmath39 of @xmath36 and its singular vector ( theorem [ mainm01 ] ) . 2 . the smallest relative singular value @xmath40 is bounded from above by @xmath41 where @xmath42 and @xmath43 are explicitly computable constants and @xmath44 is exponentially decaying in @xmath45 ( theorem [ maindist ] ) . whereas the first property allows to prove that an eigenvalue is in a certain interval , the second property allows to determine intervals in which there are no eigenvalues . both estimates together can be used to find all eigenvalues in a specified interval . we demonstrate that these matrices can be computed within a given precision and show that the singular values can be bounded from above and below using interval arithmetics . as a proof of concept we implemented our method in fortran and in mathematica . this resulted in programs that allow to compute eigenvalues rather accurately for a surface of genus @xmath16 with given fenchel - nielsen coordinates . a mathematica program was used to compute the first eigenvalues of the bolza surface with extremely high accuracy . the spectral zeta function @xmath46 is defined as the meromorphic continuation of the function @xmath47 where @xmath48 are the non - zero eigenvalues of the laplace operator repeated according to their multiplicities . the ( zeta - regularized ) spectral determinant @xmath49 is defined by @xmath50 ( zero is not a pole of the meromorphic continuation ) . since the meromorphic continuation is contructed from the full spectrum it is a priori not enough to know a finite part of the spectrum to compute the spectral determinant up to a certain accuracy . however , we show in section [ zeta ] that the selberg trace formula may be used in conjunction with the list of eigenvalues up to a certain threshold to calculate values of the spectral zeta function , in particular the casimir energy and the spectral determinant , up to a certain precision depending on that threshold . the spectral determinant for the laplace operator on hyperbolic surfaces is of particular importance . in the seminal paper @xcite osgood , phillips and sarnak showed that the spectral determinant is maximized at the hyperbolic metric in each conformal class of a given volume . thus , understanding the extremal properties of the determinant as a function on the space of metrics of fixed volume in dimension two is equivalent to the understanding of the spectral determinant as a function on the teichmller space of hyperbolic metrics . of course values of the determinant for non - hyperbolic surfaces can be computed from the value for the corresponding uniformized hyperbolic surface using the polyakov formula . finally , section [ zetaexa ] contains of collection of examples of interesting surfaces of genus two and three for which we computed the spectral determinant and the value of the spectral zeta function at the point @xmath51 . the set of examples contains the isolated surfaces of genus two with large symmetry group which are well known to be critical points for any value of the spectral zeta function or the spectral determinant ( see e.g. @xcite ) . we conjecture that the value of the spectral determinant is maximized at the bolza surface and we compute the value of the spectral determinant rather accurately . such a conjecture would imply that the global maximum of the spectral determinant as a function on the space of all metrics of fixed volume is attained at this point . the appendix contains explicit estimate for various resolvents as well as explicit estimates of the @xmath52 norm of eigenfunctions and derivatives of eigenfunctions . these estimates are needed to make the constants in our estimates explicit but they may also be interesting in their own right . eigenvalues of hyperbolic surfaces of genus two have been calculated in the physics literature by aurich and steiner in @xcite using the finite element method , and in @xcite using the boundary element method . in both cases the authors relied on the realization of the surface by geodesic octagons . it is quite interesting that the hadamard - gutzwiller model discussed in @xcite is actually the same as the bolza surface . using the group action and the decomposition into geodesics triangles ninnemann @xcite computed the eigenvalues of the regular geodesic octagon with every second side identified . the resulting surface , which he also refers to as the gutzwiller octagon , does however not coincide with the bolza surface . its fenchel nielsen coordinates are given in section [ regoct ] . there as well as in other parts of the physics literature the bolza surface is referred to as the regular octagon . the method of particular solutions is based on an article by fox , henrici and moler @xcite . it was subsequently further developed and revived by betcke and trefethen @xcite to achieve high accuracy in eigenvalue computations on domains with dirichlet boundary conditions . further versions , including domain decompositions and matching of boundary data including the normal derivative were described in @xcite and @xcite . explicit error bounds for the eigenvalues in the method of particular solutions for domains in @xmath53 were given in @xcite and further improved in by alex barnett @xcite . we would also like to refer to the latter article for further background and literature on the method of particular solutions . our estimate for the error bound does not use @xcite but is instead based on the more refined information contained in the resolvent of the laplace operator on the manifold . thus , for generic eigenfunctions it is expected to give an improvement of the order of the square root of the eigenvalues . estimates that provide such an improvement without this genericity assumption have recently been obtained for domains in @xmath54 in @xcite and @xcite . finally , for hyperbolic surfaces with cusps hejhal ( @xcite and @xcite ) introduced a method to compute maass cusp forms which is based on the group action and the expansion of the embedded eigenvalues on the cusp . after most of the work in this paper was completed we learned that booker , strmbergsson and venkatesh @xcite have recently used the pre - trace formula together with the taylor expansion of boundary data of quasi - modes to rigorously verify embedded eigenvalues for the modular surface @xmath55 . their way of certifying eigenvalues is in spirit similar to the method we use to prove that our computed eigenvalues are accurate within the error bounds . apart from the fact that their method applies to a different geometric situation ( surfaces with at least one cusp ) the method of bounding the error is different when it comes to technical detail : for example in @xcite it is estimated in terms of the @xmath52-norm of the boundary data instead of the @xmath56-norm . it would be very interesting in the future to combine these ideas . in particular a modification of our estimates applies to hyperbolic surfaces with cusps . the problem of adjusting the step size in the search for maass cusp forms on @xmath57 in such a way that no eigenvalues are missed could be tackled in this way . formulae for the spectral determinant based on the length spectrum were given in the mathematics literature by fried @xcite and by pollicott and rocha @xcite , and in the physics literature by aurich and steiner @xcite . in @xcite , exponential convergence was proved and a numerical algorithm was established for the case of surfaces of genus @xmath58 in @xmath59-fenchel nielsen coordinates without twisting . we are not able to confirm the numerical values obtained in @xcite in the three examples there but obtain quite different values . we believe that the part of the length spectrum computed in @xcite was not sufficient to obtain the correct values . in contrast to the other methods employed our approach to compute the spectral determinant and the values for the zeta function allows for explicit error estimates even if the length spectrum is unknown . the method of particular solutions is a method to approximate eigenvalues and eigenfunctions of a differential operator . we will consider here the case of the laplace operator @xmath60 acting on functions on a compact oriented riemannian manifold @xmath61 . to fix notations suppose that @xmath62 is a closed subset which is the finite union of oriented compact codimension one submanifolds @xmath63 , possibly with boundary . we will assume here for simplicity that the @xmath63 intersect at most at boundary points , i.e. @xmath64 . removing @xmath65 from @xmath61 results in an open manifold @xmath66 that is the interior of a manifold with piecewise smooth boundary . this manifold might however have corners or other singularities . we would like to define the class of functions that are `` smooth up to the boundary '' on @xmath66 . in order to do this let us be more precise about how the manifold with piecewise smooth boundary is constructed from @xmath61 and @xmath65 . the space @xmath66 equipped with the geodesic distance is not complete and we denote by @xmath67 its abstract metric completion . since @xmath61 is complete there is a natural continuous surjection @xmath68 . we define a function @xmath69 to be smooth if and only if for every point @xmath70 there exists an open neighborhood @xmath71 such that there exists a smooth function @xmath72 on @xmath61 with @xmath73 . the corresponding sheaf of smooth functions endows @xmath67 with the structure of a smooth manifold with piecewise smooth boundary . we write @xmath74 for the space of smooth functions on @xmath67 . of course an element @xmath75 can be understood as a function in @xmath76 as the boundary of @xmath67 has zero measure and the interior of @xmath67 can be identified with @xmath66 . we will now make this identification without further mention . since each @xmath63 is oriented there is a natural unit normal vector field @xmath77 to @xmath63 and a closed tubular neighborhood of @xmath63 diffeomorphic to @xmath78 $ ] . a function @xmath75 therefore has two boundary values on @xmath63 , a right boundary value @xmath79 and a left boundary value @xmath80 . both are smooth functions on @xmath63 . of course the function @xmath81 is continuous on @xmath61 if and only if for all indices @xmath82 @xmath83 similarly , @xmath84 if and only if for all @xmath82 @xmath85 here , @xmath86 denotes the right and left limits of the normal derivative of @xmath32 . let us define @xmath87 to be the function on each manifold @xmath63 given by @xmath88 and let @xmath89 be the function given by @xmath90 . let @xmath91 and @xmath92 be the corresponding functions in @xmath93 . as the boundary of @xmath65 has zero measure in @xmath65 these functions are well defined and they are smooth on each @xmath63 . let @xmath94 . the functional @xmath95 defined by @xmath96 measures the continuity of the function and its normal derivatives in different sobolev norms . here the @xmath97 norm on @xmath65 is defined as @xmath98 let now as above @xmath75 and let @xmath99 be defined by @xmath100 for @xmath101 . suppose that @xmath102 is any test function . then , by green s formula ( see e.g. prop . 4.1 in @xcite ) , @xmath103 where @xmath104 denotes the unit normal vector field along @xmath65 , @xmath105 is the measure induced by the riemannian metric on @xmath61 , and @xmath106 are the measures on @xmath107 and @xmath65 induced by the riemannian metrics there . if we think of @xmath32 as a distribution on @xmath61 , then the above simply means that we have in the distributional sense @xmath108 where @xmath109 is the dirac delta distribution on @xmath65 and @xmath110 is its normal derivative so that for a test function @xmath102 the above distributions are given by @xmath111 and @xmath112 note that the distribution @xmath113 is in @xmath114 . denote by @xmath115 its @xmath116-norm , where the @xmath97-norm is defined by @xmath117 . by the trace theorem for sobolev spaces ( see e.g. prop . 4.5 in @xcite ) the maps @xmath118 and @xmath119 extend to bounded maps from @xmath120 and @xmath121 respectively . hence , by duality , @xmath122 and @xmath123 for some constants @xmath124 . thus , there exists a constant @xmath125 depending only on the geometry of @xmath61 and @xmath65 such that @xmath126 a constant @xmath127 in this estimate can be explicitly found for a given manifold , for example by using gluing functions and the fourier transform . we will illustrate in section [ sec : mps_on_surface ] how to get a bound on @xmath127 for a given hyperbolic surface when @xmath65 consists of geodesic segments ( cf . ( [ sbound1 ] ) and ( [ sbound2 ] ) ) . the method of particular solutions relies on the following simple observation . [ main01 ] suppose that @xmath128 with @xmath129 such that @xmath130 if @xmath131 , then there is an eigenvalue of @xmath4 in the interval @xmath132.\ ] ] as before let @xmath133 then @xmath134 and by our assumptions @xmath135 now note that @xmath136 of course @xmath137 and @xmath138 as @xmath139 . from this it immediately follows that in case @xmath140 exists its operator norm is bounded from below by @xmath141 this implies the theorem since the operator norm of the resolvent at the point @xmath31 is bounded from above by the inverse of the distance of @xmath31 to the spectrum . [ remrem ] the statement of the previous theorem also applies to situations where eigenvalues might be close to each other or have high multiplicities . not to overload notation we state this here only as a remark . if there is an orthonormal set @xmath142 in @xmath12 such that for each of the @xmath32 the conditions of theorem [ main01 ] hold , then the same proof shows that there are at least @xmath45 eigenvalues ( counting multiplicities ) in a small interval around @xmath31 . to apply the method of particular solutions one constructs normalized functions @xmath32 which are eigenfunctions of the laplace operator on each of the @xmath143 with eigenvalue @xmath31 and for which @xmath144 is small . the above theorems then show that @xmath31 must be close to an eigenvalue and @xmath32 is close to an eigenfunction . we will show in the following how this can be done for oriented hyperbolic surfaces of genus @xmath16 . a hyperbolic surface is a @xmath58-dimensional orientable riemannian manifold endowed with a metric of constant negative curvature equal to @xmath145 . such surfaces can be obtained by factorizing the hyperbolic plane @xmath146 with metric @xmath147 by a discrete co - compact hyperbolic subgroup @xmath65 of @xmath148 . the isometric action of @xmath148 is given by fractional linear transformations @xmath149 another way to obtain a surface of genus @xmath16 with hyperbolic metric is to glue @xmath150 hyperbolic pair of pants . a hyperbolic pair of pants is a genus @xmath151 surface with boundary @xmath152 equipped with a metric of constant negative curvature @xmath145 such that the boundary curves are geodesics ( see figure [ pants ] ) . we will also refer to such a surface as a @xmath153-piece . -piece with boundary geodesics , width=113 ] such hyperbolic pairs of pants are , up to isometry , uniquely determined by the length of their boundary geodesics @xmath154 . any hyperbolic surface can be decomposed into pairs of pants by cutting along @xmath155 non - intersecting simple geodesics on the surface ( see figure [ genus3figure ] ) . this results in @xmath150 pairs of pants . -pieces , width=377 ] a pair of pants can be glued from a right angled geodesic octagon in the upper half space ( see figure [ octi ] ) by identifying the sides @xmath156 and @xmath157 , as well as @xmath158 and @xmath72 respectively . as indicated in the figure the geodesic octagon is constructed from two right angled geodesic hexagons . -piece glued from a right angled geodesic octagon , width=188 ] note that identification of the sides @xmath156 and @xmath157 only yields a subset of a hyperbolic cylinder so that every pair of pants can be constructed from a hyperbolic cylinder by cutting and gluing . the first ingredient in our algorithm is an estimate on the constant in the estimate ( [ const ] ) for @xmath159 in the particular case when @xmath61 is a hyperbolic surface and @xmath65 is a finite union of geodesic segments . to start let us assume that @xmath65 consists of exactly one geodesic segment . the general case of a finite union of geodesic segments will later be reduced to this case . we can assume that @xmath61 is realized as a quotient of the upper half plane @xmath160 by some co - compact hyperbolic subgroup in @xmath161 and that @xmath162 is an ( open ) fundamental domain whose boundary is a geodesic polygon . this means we can identify functions on @xmath61 with functions on @xmath160 that are invariant under the action of the group . we can also assume that the geodesic segment @xmath65 is a segment on the imaginary axis . then there is a tesselation of @xmath160 by translates @xmath163 of @xmath164 . let @xmath165 and suppose that @xmath166 is a finite number of translates of @xmath164 such that @xmath167 implies that @xmath168 is contained in the closure of @xmath169 . if @xmath170 is chosen as the smallest distance between non - adjacent sides of @xmath164 then @xmath171 would be the minimal number of fundamental domains of @xmath164 needed to cover an open neighborhood of @xmath172 ( and thus the number of neighboring fundamental domains sharing either an edge or a corner with @xmath164 ) . let @xmath173 $ ] be smooth compactly supported function which is equal to one in an open neighborhood @xmath174 of @xmath164 with support contained in @xmath175 . now suppose that @xmath157 is a distribution on @xmath61 . then , @xmath176 using @xmath177 where the inner product is with respect to the metric , we get @xmath178 of course @xmath179 since @xmath180 is invariant under the group action @xmath181 and therefore we obtain @xmath182 with @xmath183 consequently , @xmath184 to continue we use a particular coordinate system on the upper half space . namely , we identify the upper half space with @xmath185 using coordinates @xmath186 and the metric @xmath187 such that @xmath188 coincides with the imaginary axis and the coordinate system is centered at the point @xmath189 . note that @xmath186 is related to fermi coordinates @xmath190 by @xmath191 . we assume that in these coordinates @xmath65 is identified with the segment @xmath192 $ ] so that @xmath157 has the form @xmath193 moreover in these coordinates @xmath4 has the form @xmath194 where @xmath195 is the euclidean laplace operator in coordinates @xmath186 . then we have @xmath196 where @xmath197 . collecting all terms we obtain @xmath198 however , @xmath199 where @xmath200 and @xmath201 are understood as functions on @xmath202 extended by zero from @xmath203 $ ] . therefore , we finally obtain @xmath204 where the constant @xmath127 is given by @xmath205 if @xmath65 consists of several geodesic segments then of course the same inequality holds with the constant being the maximum of the constants for the individual segments . for low lying eigenvalues it is in practice easier to compute the @xmath56-norms of @xmath200 and @xmath201 . of course since the negative sobolev norms are dominated by the @xmath56-norms the above gives estimates of @xmath206 in terms of the @xmath56-norms of @xmath200 and @xmath201 . however , in this case a slightly better constant can be obtained using theorems [ app1 ] and [ app2 ] . indeed , equation [ sob ] means by duality that @xmath207 where @xmath208 is the laplace operator on the hyperbolic plane . by theorem [ app1 ] and [ app2 ] we have @xmath209 where @xmath210 and @xmath211 are the constants in these theorems . for a surface of genus @xmath58 we use a right angled 12-gon and thus have @xmath212 . let @xmath170 be the minimum of the distance between non - adjacent sides . for each side @xmath213 we use a cutoff function @xmath214 , where @xmath215 is the function @xmath216 , \end{cases}\ ] ] and @xmath217 are fermi - coordinates with respect to the infinite extension of the side @xmath213 oriented in such a way that the region @xmath218 does not intersect the polygon . note that @xmath219 and @xmath220 . with @xmath221 we then get @xmath222 where we used the well known formula @xmath223 and the fact that near a fixed point all but at most two of the functions @xmath224 are constant . since @xmath225 is not smooth we can not directly use it in our estimate . however , we can use a suitable regularization @xmath226 and the fact the the second distributional derivatives of @xmath225 are in @xmath52 to show that the estimate above still holds with the function @xmath225 . collecting all terms we get @xmath227 since @xmath228 this further simplifies into the easy to use estimate @xmath229 this estimate improves by a factor of @xmath230 if one is willing to believe the numerical values computed at the end of appendix [ constest ] . let @xmath231 . then the hyperbolic cylinder @xmath232 is the non - compact manifold obtained by factorizing the upper half space by the subgroup of @xmath148 generated by the element @xmath233 . we will use fermi coordinates @xmath190 which are related to the usual coordinates by @xmath234 for @xmath165 denote by @xmath235 the finite hyperbolic cylinder @xmath236 for @xmath237 let @xmath238 be the space of functions @xmath180 in @xmath239 that satisfy @xmath240 note that @xmath241 has a fourier expansion of the form @xmath242 where @xmath243 satisfy @xmath244 and have initial data @xmath245 and the functions @xmath246 satisfy @xmath247 and have initial data @xmath248 these functions are of the form @xmath249 , where @xmath250 solves the ordinary differential equation @xmath251 with the corresponding initial conditions . a fundamental system of ( non - normalized ) solutions of this equation , consisting of an even and an odd function , can be given explicitly in terms of hypergeometric functions @xmath252 where @xmath253 ( see @xcite , where these functions are analysed ) . normalization gives the corresponding solutions to the initial value problems . denote by @xmath254 be @xmath255 dimensional subspace of @xmath256 spanned by @xmath257 and @xmath246 with @xmath258 . given @xmath259 we can truncate the fourier expansion to obtain an element @xmath260 given by @xmath261 the @xmath262-norm on the cylinder @xmath263 is defined as @xmath264 the following result is crucial for our approximation of eigenfunctions . [ bestim ] suppose @xmath265 is bounded on @xmath232 . let @xmath165 and suppose that @xmath266 is an integer such that @xmath267 . then , @xmath268 where @xmath269 , @xmath270 @xmath271 and @xmath272 both @xmath273 and @xmath274 are decreasing exponentially fast in @xmath266 for fixed @xmath31 and give very good bounds for reasonably large @xmath266 ( see fig . [ betas ] ) . on a logarithmic scale for @xmath275 , @xmath276 and @xmath277 . the corresponding cylinder covers a fundamental domain for the bolza surface.,width=377 ] by rescaling we can assume in the proof that @xmath278 . of course @xmath279 has a convergent expansion into a fourier series @xmath280 which converges uniformly on compact sets . both @xmath257 and @xmath246 are of the form @xmath281 where @xmath282 solves the ordinary differential equation ( [ dequ ] ) with initial conditions either @xmath283 or @xmath284 respectively . thus , @xmath282 is either even or odd with respect to the reflection @xmath285 . substituting @xmath286 we write @xmath287 . let @xmath288 and define @xmath289 so that @xmath290 is non - negative on the interval @xmath291 $ ] . then , by lemma [ growth1 ] , we have for all @xmath292 and @xmath293 : @xmath294 where @xmath295 this integral can be computed explicitly and gives the expression in the theorem . therefore @xmath296 since the @xmath297 is bounded by @xmath298 we have @xmath299 and @xmath300 . for any @xmath218 . with @xmath301 the above bound then implies that for any @xmath302 $ ] both @xmath303 and @xmath304 are smaller equal than @xmath305 and @xmath306 and @xmath307 are smaller equal than @xmath308 summing the fourier series one obtains on @xmath309 the bounds @xmath310 , @xmath311 and @xmath312 . suppose that @xmath61 is a surface of genus @xmath16 . then @xmath61 can be decomposed into @xmath313 and @xmath153 pieces ( pairs of pants ) . each of these pieces can be cut open along a shortest geodesic that connects two different boundary components to obtain a surface that can be identified with a closed subset of a hyperbolic cylinder ( see fig . [ cyl ] ) . -piece that was cut open along two geodesic segments covered by a hyperbolic cylinder . the hyperbolic cylinder is obtained from the shaded region by identification of the two dashed circles . the inscribed finite cylinder is double shaded.,width=377 ] in this way we obtain a collection @xmath314 of closed subsets of hyperbolic cylinders @xmath315 with geodesic boundary components in such a way that the defining unique simple closed geodesic @xmath316 on @xmath315 is contained in @xmath143 . note that we use disjoint cylinders here , i.e. we use a different copy of @xmath315 even if the geodesics have the same length and correspond to the same model in the upper half space . as @xmath316 is a simple closed geodesic in @xmath61 the subgroup it generates in the fundamental group will define a covering map @xmath317 . therefore , every eigenfunction @xmath32 on @xmath61 lifts to a bounded smooth function on @xmath315 which is an eigenfunction of the laplace operator on @xmath315 . the surface @xmath61 can of course be constructed from the @xmath143 by glueing , i.e. by identifying the different boundary geodesics . suppose that we have a collection @xmath318 of geodesic segments and a collection @xmath319 so that the geodesic segment @xmath63 is identified with the segment @xmath320 . thus , @xmath321 can be thought of as a finite union of geodesic segments on @xmath61 and @xmath66 is the interior of the disjoint union @xmath322 of the @xmath143 . this is exactly the setting described in section [ mps ] where @xmath67 gets identified with @xmath322 and the segments @xmath323 and @xmath324 form the boundary of @xmath67 . thus , for a function @xmath325 which is smooth up to the boundary the boundary values at @xmath323 and @xmath324 represent the two different boundary values it has along @xmath63 if we think of it as a function on @xmath61 . on each segment @xmath63 we choose a set @xmath326 consisting of an even number of equidistant points such that @xmath327 is the origin of the segment and @xmath328 is the endpoint . let @xmath329 be the corresponding points on @xmath324 . we denote by @xmath330 the distance between neighbouring points on the segment @xmath63 and by @xmath331 the maximum of @xmath332 . let @xmath333 be a fixed integer and suppose @xmath1 . of course on each cylinder @xmath315 and for each integer @xmath334 such that @xmath335 there exists two linearly independent functions on this cylinder which are both eigenfunctions of the laplace operator with eigenvalue @xmath31 and which are of the form @xmath336 where @xmath337 are fermi coordinates on the corresponding cylinder . we are choosing the basis functions orthonormal on the largest subset of @xmath143 which is invariant under the rotation symmetry of the cylinder ( for a @xmath153-piece this will be one half of a cut - off hyperbolic cylinder and corresponds to the double shaded region in fig . [ cyl ] ) . since such a subset will always have the form @xmath338 \times s^1 $ ] this can be achieved by choosing @xmath339 orthonormal on this interval . we extend these functions by zero to the disjoint union of all the cylinders and think of the collection of all these functions as a basis @xmath340 in the space @xmath341 we will use this basis to identify vectors in @xmath342 with functions on the disjoint union of the cylinders , but also with @xmath52 functions on @xmath61 , by restricting the functions to the pieces @xmath143 . for @xmath343 let @xmath344 be the corresponding function . for each geodesic segment @xmath63 we form the matrices @xmath345 where @xmath104 is the outward normal unit vector field along the geodesic segment , @xmath346 the normalized tangential vector field , and @xmath347 are the coefficients in the composite simpson rule , i.e. @xmath348 . next arrange @xmath349 and @xmath350 into one matrix @xmath351 , and add also the tangential derivatives @xmath352 . all the matrices involved map into the same space and we understand the direct sum here as a direct sum of linear maps mapping into @xmath342 . thus , the matrix @xmath353 is obtained from by attaching all the rows of @xmath354 as rows to the matrix @xmath355 ( similar to the matlab notation @xmath356 $ ] ) . similarly , but with one change of sign @xmath357 and again @xmath358 , @xmath359 . now define @xmath360 as well as @xmath361 , @xmath362 and @xmath363 . thus , @xmath364 as compared to @xmath365 contains also the tangential derivatives . for each @xmath343 the vector @xmath366 contains the discretization of the jump of boundary data of @xmath344 across @xmath65 whereas @xmath367 contains the discretization of the boundary data itself . we will see below that the @xmath56 norm of an eigenfunction on @xmath61 can be bounded from above and below by the @xmath56-norm of the boundary data and thus the norm of @xmath367 will serve as a normalization in our method . to be more precise , the matrices are chosen such that the norms @xmath368 and @xmath369 are the approximations of @xmath370 and of the @xmath56 norm of the boundary data of @xmath344 using the simpson rule . the error of composite simpson integration of a function @xmath72 on an interval @xmath371 $ ] with step size @xmath331 is bounded from above by @xmath372 } \| g^{(4)}(\xi ) \|$ ] ( see e.g. @xcite , p. 133 ) . thus , we have @xmath373 where @xmath374 denotes the tangential derivative . moreover , as we chose the basis functions orthonormal with respect to the @xmath56-inner product of a subset of @xmath61 , we have @xmath375 assume now that @xmath31 is an eigenvalue of the laplace operator and suppose that @xmath32 is an @xmath56-normalized eigenfunction . we may truncate the fourier expansion of this eigenfunction on each of the cylinders to obtain a function @xmath376 which then corresponds to a fixed vector @xmath377 such that @xmath378 . we choose @xmath266 large enough such that theorem [ bestim ] applies and from the fact that @xmath32 is bounded by an explicitly computable constant ( see appendix , corollary [ lbound ] ) the following estimates are explicit @xmath379 the error of the composite simpson integration of a function @xmath72 on an interval @xmath371 $ ] with step size @xmath331 is also bounded by @xmath380 } \| g'(\xi ) \|$ ] as one can easily see by replacing @xmath72 in each sub - interval @xmath381 $ ] of length @xmath382 by the constant function @xmath383 . therefore ) because a bound on the first derivative of eigenfunctions is easier to obtain explicitly and we do not need good accuracy for this bound ] , we obtain @xmath384 from the bound ( [ l2bound ] ) of the appendix and the fact that @xmath32 is @xmath56-normalized we conclude that @xmath385 where @xmath44 is explicitly computable and depends only on @xmath31 and the geometry of @xmath61 and @xmath65 . using the bound @xmath52 and the @xmath262 bound in the appendix of eigenfunctions we can compute constants @xmath386 and @xmath387 such that for all @xmath388 and @xmath389 we have @xmath390 note that this is a very crude estimate and we have used only the linear error estimate in simpson s rule . this was done so that we can use the explicit estimates that we proved in the appendix . the critical constants @xmath387 and @xmath386 guarantee that the numerical integration using the simpson rule together with the approximation of @xmath32 by the truncated function together yield a relative error of not more than 50% . for the method it will be enough to show that @xmath369 is bounded from above so that this estimate is sufficient . let @xmath391 be the smallest singular value of @xmath392 and @xmath393 the smallest generalized singular value ( see e.g. @xcite or also @xcite in the context of mps ) of the pair @xmath394 . here the generalized singular values of @xmath395 with respect to @xmath396 are defined as the singular values of @xmath397 as an operator from @xmath398 to @xmath399 where the euclidean norm @xmath400 on @xmath398 is replaced by the norm @xmath401 defined by @xmath402 . note that @xmath403 if @xmath266 is large enough , so that the assumptions of theorem [ bestim ] are satisfied , we have @xmath404 in case @xmath31 is an eigenvalue . this follows immediately from the fact that an exact eigenfunction satisfies the boundary conditions . let as usual @xmath253 with @xmath405 and define @xmath406 and @xmath407 . assume that @xmath408 is not an eigenvalue and let @xmath409 and @xmath410 defined as above . for each @xmath411 we choose a hyperbolic cylinder @xmath309 with @xmath170 large enough such that @xmath412 and such that @xmath143 is relatively compact in the interior of @xmath309 , so that the distance between @xmath413 and @xmath143 greater than a positive number @xmath414 . for the truncated eigenfunction @xmath376 consider the boundary data @xmath415 on @xmath413 . using the resolvent kernel of the laplace operator on the hyperbolic cylinder ( see appendix [ resestcly ] ) we can reconstruct @xmath376 from its boundary data @xmath416 now define @xmath417 then @xmath418 using lemma [ greenestcylder ] we obtain @xmath419 of course the @xmath56-norm of the boundary data is bounded from above by the @xmath56-norm of the boundary data of the exact eigenfunction @xmath32 so that we obtain the bound @xmath420 [ maindist ] with the notation above we have for all @xmath388 and @xmath389 @xmath421 if @xmath31 is an eigenvalue with normalized eigenfunction @xmath32 . here @xmath422 and @xmath423 are bounded by explicit constants ( corollary [ lbound ] and [ lboundzwei ] ) . moreover , by theorem [ main01 ] and the inequalities ( [ simpson01 ] ) and ( [ geq01 ] ) we have : [ mainm01 ] suppose that @xmath343 is a unit vector and @xmath424 and let @xmath425 where @xmath426 are the constants in section [ s:4.1 ] . then , there exists an eigenvalue in the interval @xmath427 $ ] . in particular the estimate holds if @xmath428 and @xmath377 is a corresponding normalized singular vector . the term @xmath429 can be minimized as another singular value problem , using the bounds on the derivatives of the basis functions that follow from the differential equation together with the lemmata [ oni1 ] and [ oni3 ] . this results in a matrix , whose singular value is precisely the error estimate . actual computations show that in practice the major contribution of the error comes from the first term . an analog of theorem [ maindist ] holds for the @xmath3-th relative singular values @xmath430 with respect to @xmath431 . in this case the term @xmath432 gets replaced by the maximum of @xmath433 , where @xmath434 correspond to @xmath3 pairwise orthogonal eigenvalues . we do not state this here as we want to keep the exposition simple . in the same way theorem [ mainm01 ] applies to situations with multiplicities and eigenvalues that are close to each other ( see remark [ remrem ] ) . since the fundamental solutions of the differential equation ( [ dequ ] ) satisfy the bounds of lemmata [ oni1 ] and [ oni3 ] their higher derivatives can also be bounded explicitly simply by using the differential equation . we use coordinates @xmath435 and @xmath436 on the hyperbolic cylinders , so that the differential equation is of the form @xmath437 using subdivision into smaller intervals and the taylor expansion around the boundary points of these intervals yields an approximation @xmath438 to the exact solution by piecewise defined polynomial splines . using the bounds on the higher derivatives the error in the @xmath262-norm can be explicitly bounded from above . since the measure on the hyperbolic cylinder is equal to @xmath439 in these coordinates , normalization of the basis functions can be done within this error , simply by integrating these piecewise polynomials . in this way , the matrices @xmath392 , @xmath397 and @xmath440 can be computed with explicit error bounds . the generalized singular value decomposition can then be done using any method from linear algebra . once a numerical generalized singular value decomposition has been obtained the error of the singular values can be estimated from above as follows . suppose that @xmath441 is a numerically obtained generalized singular value decomposition of @xmath442 with respect to @xmath443 . thus , @xmath444 is diagonal and @xmath445 is unitary within a certain precision . moreover , @xmath446 is unitary with respect to the inner product induced by @xmath443 , again within a certain precision . thus , @xmath447 and @xmath448 , and @xmath449 , where these three numbers reflect the errors . all three numbers can be estimated from above using interval arithmetics for example by bounding the operator norm by the hilbert - schmidt norm . since the matrices @xmath450 and @xmath451 are unitary in the respective inner products the generalized singular values of @xmath452 are exactly the diagonal entries of @xmath444 . by the characterization of generalized singular values of @xmath442 as the norm distance of @xmath442 to the rank @xmath45 operators it follows that the error of the generalized singular values of @xmath442 bounded by @xmath453 using functional calculus a longer calculation shows that this is bounded by @xmath454 where @xmath455 is the smallest singular value of @xmath443 . here we assumed that @xmath456 . note that the smallest singular value of @xmath443 can also be estimated from below using exactly the same procedure for the ordinary singular value decomposition of @xmath443 . thus , interval arithmetics can be applied to obtain a lower bound for the generalized singular values of @xmath442 with respect to @xmath443 . an interval @xmath457 is then tested for eigenvalues by computing a lower bound for the value of @xmath458 for a discrete set of points @xmath168 in @xmath457 . by theorem [ maindist ] for large enough @xmath266 the distance between the points can be chosen such that eigenvalues can occur only near points where the lower bound is small enough . for a fixed @xmath459 this algorithm then will normally find a discrete set of points @xmath460 such that eigenvalue can occur only in @xmath0-neighborhoods of these points . the presence of eigenvalues and their multiplicities can then be tested using the singular values of @xmath392 . an upper bound for small singular values can easily be obtained by first finding the singular value decomposition numerically and then using the numerically obtained singular vectors @xmath377 to compute @xmath368 . since the derivatives of the basis functions can be bounded using the differential equation ( [ dequ ] ) and the bounds of lemmata [ oni1 ] and [ oni3 ] the error terms in theorem [ mainm01 ] can directly be estimated from above . using interval arithmetics one can therefore obtain rigorous interval inclusions for the eigenvalues . once the presence of a single eigenvalue is established in a small interval theorem [ maindist ] can also be used to establish an error bound by narrowing the interval in which the eigenvalue can be located . the analysis of rigorous error bounds can be simplified considerably by replacing the exact basis functions with the piecewise defined polynomial basis functions themselves . these are of course orthogonal on any cylinder , thus normalization involves only integration of piecewise defined polynomials which can be done explicitly . thus , the quasi - mode becomes a finite linear combination of products of piecewise defined polynomials in @xmath461 and fourier modes in @xmath436 . the so constructed function does not satisfy the equation @xmath462 but instead @xmath463 , where due to the nature of the differential equation @xmath464 is again constructed out of piecewise defined polynomials and fourier modes . its @xmath56-norm can very easily be estimated from above by integrating over the smallest cylinders that contain @xmath143 . the fourth derivatives in the error of simpson s rule in theorem [ mainm01 ] can also be estimated more directly in this case as a bound can easily obtained from the coefficients of the polynomial . we have implemented our method in three different programs . the first two programs are fortran programs . one works solely for genus @xmath58 surfaces and takes as coordinates @xmath59-fenchel - nielsen coordinates . the second program takes as an input a labeled graph ( encoding the way the surface is glued from @xmath153-pieces ) and the fenchel - nielsen coordinates . both programs then follow the algorithm described in this article . the basis functions are computed for each value of @xmath31 by numerically solving the differential equation ( [ dequ ] ) using the dlsoda routine from odepack ( see @xcite ) . normalization of the basis functions is achieved by numerical integration of the solution of ( [ dequ ] ) . in order to compute the singular values @xmath391 we use lapack routines . the generalized singular values @xmath393 are computed using @xmath465-decomposition . a search algorithm then looks for the small minima of @xmath393 as a function of @xmath31 . a small enough step size guarantees that no eigenvalues are missed in the process . the other singular values are also necessary in the search : a small second singular value of @xmath392 implies that either the eigenvalue has multiplicity greater than one , or another eigenvalue is very close . our method can not distinguish between multiplicities and close eigenvalues if they can not be separated within the given precision . both programs achieve accuracies close to machine precision for the low lying eigenvalues when the surface is far enough from the boundary of teichmller space ( so that the dlsoda algorithm can compute the basis functions accurately enough ) . the code for the genus two program is available to the scientific community under gplv3 and can be obtained from http://www-staff.lboro.ac.uk/~maas3/hyperbolic-surfaces/hypermodes.html . another program was written in mathematica . it uses taylor s method on intervals of size of order @xmath466 in order to solve the differential equation ( [ dequ ] ) and compute the matrices @xmath397 and @xmath440 with arbitrarily high precision . the linear algebra such as @xmath465-decomposition and search for the minimum is done within mathematica . the achieved accuracy depends on the computing time and the machine precision used . we were able to compute the first 70 digits of the first eigenvalue for the bolza surface ( see section [ bolzasection ] ) using this program . to demonstrate the power of our method we modified the genus two fortran and mathematica codes to give rigorous error estimates for the eigenvalues . as we described in the previous section we use the taylor series method to solve the differential equation ( [ dequ ] ) within a given accuracy and then follow the analysis described in section [ algozwei ] . the corresponding eigenvalue inclusions can be considered rigorous if one carefully includes rounding errors in the analysis and is willing to trust the implementation of basic mathematical functions in fortran and mathematica . in order to demonstrate our method we consider a special surface of genus @xmath58 . the so - called bolza surface has fenchel - nielsen @xmath59-coordinates @xmath467 this surface is known to maximize the length of the systole in genus @xmath58 and it also maximizes the order of the symmetry group ( see e.g. @xcite ) . the first eigenvalue on the bolza surface was estimated by jenni @xcite remarkably accurately to be in the interval @xmath468 $ ] . aurich and steiner @xcite ( see also @xcite)give the value @xmath469 which was obtained using the finite element method . figure [ sigbolza ] shows a plot of the singular value after @xmath465-decomposition ( @xmath470 , @xmath471 ) of the matrix @xmath472 and projection onto the first direct summand . note that at the minima the function is very small and order @xmath473 . [ sigbolza ] -decomposition for the bolza surface as a function of @xmath31.,title="fig:",width=415 ] the search algorithm in the fortran program finds the first eigenvalue within @xmath474 digits of precision . using multiple precision in mathematica we obtained the value . @xmath475 where we believe that all digits are correct . the rigorous implementation that uses very simple bounds in its current form gives a rigorous error bound of @xmath476 ( which can be improved by simply investing more computing time ) . we believe that @xmath477 is important as we conjecture that this is the maximal value of the first eigenvalue of the laplace operator for constant negative curvature @xmath145 in genus @xmath58 . this conjecture is supported by our calculations in teichmller space ( the corresponding analysis will be discussed elsewhere ) and is in line with the findings in ( @xcite ) , where it is shown that the maximal value of the first eigenvalue of the laplace operator with respect to any metric fixing the volume is attained for a singular metric on the bolza surface ( @xcite ) . the spectral zeta function @xmath46 is defined as the meromorphic continuation of the function @xmath47 where @xmath478 are the eigenvalues repeated according to their multiplicities . the above sum converges for @xmath479 . as usual a meromorphic continuation to the whole complex plane is constructed using the mellin transform @xmath480 and the fact that the heat kernel has an asymptotic expansion , by splitting the above integral into an integral over @xmath481 $ ] and one over @xmath482 . from the heat expansion it follows that zero is not a pole , which allows one to define the zeta - regularized determinant @xmath483 of the laplace operator by @xmath484 in the examples we will also consider the value of the zeta function at the point @xmath51 . in physics this value is often referred to as the casimir energy or vacuum energy . for our purposes we use the following splitting @xmath485 where @xmath459 . the selberg trace formula applied to the heat trace ( see e.g. equ . ( 193 ) in @xcite ) is @xmath486 where the second sum in the second term is over the set of primitive closed geodesics @xmath487 . splitting the integral and using the standard analytic continuation gives for @xmath488 @xmath489 where @xmath490 where @xmath491 the @xmath492 are given by @xmath493 and @xmath494 denotes the incomplete gamma function @xmath495 differentiation results in the following formula for the spectral determinant . @xmath496 where @xmath497 and @xmath498 is the generalized exponential integral which equals @xmath499 . all the integrals have analytic integrands and can be truncated with exponentially small error . they can therefore be evaluated to high accuracy using numerical integration . for fixed @xmath500 and @xmath459 not too small the sums over the eigenvalues converge rather quickly and therefore @xmath501 and @xmath502 can be computed quite accurately from the first eigenvalues only . the error made by summing only over finitely many eigenvalues can be estimated directly by lemma [ heattrb ] . for not too large @xmath459 the sums over the length spectrum are also very fast convergent . as they are exponentially decaying as @xmath0 tends to zero one can either neglect this contribution by choosing @xmath0 small enough or one can compute them using a finite part of the length spectrum . the error introduced in this way is also exponentially decreasing as @xmath503 and it can be explicitly estimated using the selberg trace formula as follows . if the length of the shortest closed geodesic is @xmath170 the second term in the selberg trace formula equ . ( [ stf ] ) is for all @xmath504 bounded by @xmath505 and the heat trace for a fixed @xmath506 can be estimated by lemma [ heattrb ] . therefore , we have for all @xmath507 @xmath508 and the right hand side converges exponentially fast to zero as @xmath503 . an analogous formula holds for the error if elements in the length spectrum up to length @xmath170 are taken into account . of course the short geodesics dominate the contribution from the length spectrum . note however that as a consequence of the collar theorem there are at most @xmath509 closed geodesics of length smaller that @xmath510 on a surface of genus @xmath511 ( see e.g. @xcite ) . in this section we compute the values of the casimir energy and the spectral determinant for the three isolated surfaces of genus two with large symmetry group as well as for other examples of genus two and genus three surfaces that have been treated in the literature before . all fenchel - nielsen coordinates are given in @xmath59-form , i.e. the @xmath512-valent graph describing this decomposition consists of two vertices that are connected by three edges . for all the surfaces we first used the fortran double precision implementation of our program to generate a list of eigenvalues . even though our algorithm allows us to choose the stepsize in such a way that we are guaranteed not to miss an eigenvalue the current implementation does not achieve that automatically . in our computations we have chosen to check completeness of the list of eigenvalues by using the selberg trace formula applied to the heat trace . this is possible as the absolute value of the contribution of the length spectrum in selberg s trace formula is bounded from above by @xmath513 , defined in equ . ( [ errheattr ] ) , so that missed eigenvalues can easily be detected comparing the heat trace computed from the selberg trace formula . note that unlike heuristic checks using weyl s law this gives a rigorous way to check completeness up to a certain value . the data files for all the examples below can be found at http://www-staff.lboro.ac.uk/~maas3/publications/eigdata/datafile.html . the bolza surface has symmetry group @xmath514 . its order , @xmath515 , thus maximizes the order of the symmetry group amongst all genus @xmath58 hyperbolic surfaces . the bolza surface is referred to as the regular octagon ( and sometimes as the hadamard - gutzwiller model ) in the physics literature as it can be obtained from a regular octagon in hyperbolic space by identifying opposite sides . its fenchel - nielsen coordinates are @xmath516 we computed the spectral determinant as well as the casimir energy @xmath517 and obtained @xmath518 where we believe that all decimal places are correct . this accuracy was obtained by computing the first @xmath519 eigenvalues with a @xmath520 digit precision , the first @xmath521 eigenvalues in quad precision and subsequently by using the simple length spectrum up to @xmath522 . we relied on the multiplicities of the length spectrum as computed in @xcite . note that the spectral determinant and the casimir energy can both be computed accurately within 5 decimal places without taking the length spectrum into account at all , using the eigenvalues only . completeness of the list of the first @xmath521 eigenvalues was checked using a list of @xmath523 computed eigenvalues . to compute the first @xmath521 eigenvalues of the bolza surface to a precision of @xmath524 digits about @xmath525 @xmath31-evaluations of generalized singular value decomposition were needed . this took about @xmath519 minutes on a @xmath526 ghz intel core i5 quad core processor ( where parallelization was used ) . [ zetaplot ] as a function of @xmath500 for the bolza surface , title="fig:",width=415 ] we conjecture that the spectral determinant is maximized in genus @xmath58 for the bolza surface and we therefore believe that the above number is important . our computations show that bolza surface as a point in teichmller space is indeed a local maximum . note that for the bolza surface is known to be a critical point for the spectral determinant . this surface has fenchel - nielsen coordinates ( see @xcite , sec . 3.5 ) @xmath528 it is also known to be a critical point for any modular invariant such as the casimir energy or the spectral determinant . computing the first @xmath521 eigenvalues with double precision and neglecting the contribution from the length spectrum other than from the @xmath524 primitive closed geodesics of length @xmath529 we obtain @xmath530 this surface has fenchel - nielsen coordinates ( see @xcite , sec . 3.5 ) @xmath532 it is also known to be a critical point for any modular invariant such as the casimir energy or the spectral determinant . computing the first @xmath521 eigenvalues with double precision and neglecting the contribution from the length spectrum we obtain @xmath533 this is a surface obtained from a regular hyperbolic octagon in the upper half space using a side identification different from that for the bolza surface ( see @xcite where also the low lying spectrum of this surface was computed ) . the fenchel - nielsen coordinates of this surface are given by @xmath534 computing the first @xmath521 eigenvalues with double precision and neglecting the contribution from the length spectrum we obtain @xmath535 eigenvalue statistics of a certain generic octagon was investigated in @xcite . the fenchel - nielsen parameters of this surface are roughly ( see @xcite ) . @xmath536 computing the first @xmath521 eigenvalues with double precision and neglecting the contribution from the length spectrum we obtain @xmath537 these are the examples treated in @xcite , where approximate values for their spectral determinants are given . the first example has fenchel - nielsen coordinates @xmath538 computing the first @xmath521 eigenvalues with double precision and taking into acount only the short primitive geodesics in the length spectrum we obtain @xmath539 the second example has fenchel - nielsen coordinates @xmath540 computing the first @xmath521 eigenvalues with double precision and taking into acount only the short primitive geodesics in the length spectrum we obtain @xmath541 the third example has fenchel - nielsen coordinates @xmath542 computing the first @xmath521 eigenvalues with double precision and taking into acount only the short primitive geodesics in the length spectrum we obtain @xmath543 the values do not agree with the values obtained in @xcite . we believe that not enough lengths of primitive geodesics were taken into account there . the following example has genus three and is a double cover of the bolza surface . it is obtained by cutting the two copies of the bolza surface open along the two geodesics of length @xmath544 used in its pants decomposition ( see section [ bolz ] ) the two resulting surfaces of type @xmath545 are then glued by identifying each boundary with the the corresponding boundary on the other copy of the surface . we computed the first @xmath546 eigenvalues on this surface and obtained @xmath547 for this accuracy the contribution of the length spectrum can be neglected . this section deals with growth estimates and bounds on the function @xmath548 that satisfies the differential equation @xmath549 with initial conditions @xmath550 substituting @xmath286 this can also be written as @xmath551 where @xmath552 satisfies @xmath553 and @xmath554 . the derivatives in different coordinates are related by @xmath555 . now define a potential @xmath556 [ oni3 ] suppose that @xmath1 and @xmath557 does not vanish on an interval @xmath558 . then the function @xmath559 is non - increasing on @xmath457 . simple differentiation gives @xmath560 which implies the statement . [ growth1 ] suppose that @xmath1 and @xmath561 on some interval @xmath562 \subset [ 0,\pi/2)$ ] and assume @xmath563 and @xmath564 . define @xmath565 then for all @xmath566 $ ] we have @xmath567 and @xmath568 since every solution can be obtained as a sum of one with @xmath569 and one with @xmath570 we can assume without loss of generality that either @xmath569 or @xmath570 . we also assume that @xmath571 is non - zero . define @xmath572 then @xmath573 and @xmath571 have the same initial data at @xmath574 . it follows from the differential equation and the assumption @xmath563 and @xmath564 that @xmath575 and @xmath576 are positive on some interval @xmath577 . moreover , where @xmath578 and @xmath579 the function @xmath571 is convex and therefore , @xmath575 and @xmath576 are non - decreasing functions on @xmath562 $ ] . in particular @xmath580 on @xmath581 $ ] . the differential equation implies @xmath582 the right hand side of this is non - negative as on can easily see by direct calculation that @xmath583 where @xmath584 is either @xmath585 or @xmath586 depending on the inital conditions . thus , @xmath587 on @xmath581 $ ] and consequently @xmath588 . the equation also implies that @xmath589 is non - decreasing which gives the inequality for the derivatives . the following lemma is a direct consequence of gronwall s inequality applied to the first order system @xmath590 [ oni2 ] suppose that @xmath1 , @xmath591 and @xmath592 on some interval @xmath562 \subset [ 0,\pi/2)$ ] then for all @xmath566 $ ] we have @xmath593 for intervals @xmath594 where @xmath557 is positive this can be slightly refined . [ oni1 ] suppose that @xmath1 and @xmath595 on some interval @xmath596 then for all @xmath597 we have @xmath598 the equation is equivalent to the system @xmath599 the operator norm of the family of matrices in this system at the point @xmath600 is given by @xmath601 which is bounded from below by @xmath602 given that @xmath603 . the lemma now follows immediately from gronwall s inequality . let @xmath604 be the laplace operator on the upper half space equipped with the poincare metric . then the integral kernel @xmath605 of the resolvent @xmath606 is given by @xmath607 here @xmath608 is the legendre q - function . using the point pair invariant @xmath609 this can be expressed as @xmath610 where @xmath611 is given by an integral ( see e.g. @xcite ) @xmath612 this formula can be used to construct a meromorphic continuation of the resolvent as a function of @xmath500 to the whole complex plane . we will need the following estimates on this function all of which follow right from the above integral formula . [ lem1 ] suppose that @xmath613 . then @xmath614 [ lem2 ] suppose that @xmath615 . then @xmath616 a similar estimate as in the above lemma also holds for @xmath617 ( c.f . @xcite , lemma 1.7 , where however ( 1.48 ) seems to be inaccurate ) [ lem3 ] suppose that @xmath613 . then , there are constants @xmath618 , depending only on @xmath619 , and a constant @xmath620 , depending only on @xmath500 , such that @xmath621 moreover , @xmath620 can be bounded by @xmath622 , where @xmath623 depends only on @xmath619 . in the important case @xmath624 one can choose @xmath625 , @xmath626 and @xmath627 , although these choices are not optimal . the statement that @xmath620 can be bounded by @xmath622 is not optimal either for large values of @xmath628 , where cancellations determine the asymptotic behaviour as @xmath629 . [ app1 ] let @xmath65 be a geodesic segment in the upper half space and let @xmath190 be fermi coordinates with respect to the infinite extension of @xmath65 . suppose a distribution @xmath630 is defined as @xmath631 where @xmath632 . then @xmath633 where the constant @xmath210 is given by ( [ const1 ] ) and satisfies @xmath634 . since the laplace operator commutes with the @xmath148-action we can assume without loss of generality that the geodesic segment is part of the imaginary axis . the function @xmath635 can be computed by convolving @xmath630 with the integral kernel @xmath605 for @xmath636 . we use fermi coordinates @xmath190 so that the integral kernel is given by @xmath637 . thus , @xmath638 since this is a convolution the generalized young inequality applies so that @xmath639 the value of the point pair invariant @xmath640 at @xmath641 in these coordinates is given by @xmath642 so that the inequality follows with @xmath643 the integral over @xmath436 can be estimated for @xmath644 as @xmath645 for @xmath646 we split the integral over @xmath436 into two parts . one with @xmath647 and one with @xmath648 . the former gives @xmath649 and the latter @xmath650 squaring and integrating finally gives @xmath651 where @xmath652 [ app2 ] let @xmath65 be a geodesic segment in the upper half space suppose @xmath190 are fermi coordinates with respect to the infinite extension of @xmath65 . suppose a distribution @xmath653 is defined as @xmath654 then @xmath655 where the constant @xmath211 is given by ( [ const2 ] ) and satisfies @xmath656 as in the proof of the previous theorem we can assume without loss of generality that the geodesic segment is part of the imaginary axis . as before we use fermi coordinates @xmath190 so that the integral kernel is given by @xmath637 . thus , @xmath657 now by the chain rule @xmath658 again , by the generalized young inequality @xmath659 thus , @xmath655 where @xmath660 splitting the integral into a region @xmath644 and @xmath646 and using the estimates [ lem1 ] and [ lem2 ] one obtains @xmath661 the numerical value of the right hand side of this inequality is about @xmath662 . both constants @xmath210 and @xmath211 can be computed using numerical integration and their values are about @xmath663 where we believe that at least two digits are correct . the corresponding estimates are non - rigorous however . we will need the following bounds on the derivative of @xmath664 with respect to @xmath500 . [ lemc1 ] let @xmath611 be the function defined in the previous section . then , for any @xmath665 we have the estimates @xmath666 where @xmath667 and @xmath668 . in particular @xmath669 and @xmath670 . simple differentiation shows that with @xmath671 : @xmath672 which yields the first estimate . the second estimate and the third estimate are obtained in a similar manner from @xmath673 and @xmath674 near zero one obtains similarly . [ lemc2 ] let @xmath611 be the function defined in the previous section . then , for any @xmath665 we have the estimates @xmath675 where @xmath668 . in the important case @xmath624 one obtains @xmath676 the latter constant can numerically be calculated and is approximately given by @xmath677 let @xmath678 be the hyperbolic cylinder obtained by factorizing the upper half space by the subgroup of @xmath148 generated by the element @xmath233 . we use fermi - coordinates @xmath190 which identifies the cylinder with @xmath679 . the integral kernel of the resolvent of the laplace operator @xmath680 can be constructed from the resolvent of @xmath208 by averaging over the group : let @xmath681 and @xmath682 be distinct points on the hyperbolic cylinder and denote by @xmath683 the integral kernel of @xmath684 . then the integral kernel @xmath685 of @xmath686 is given by @xmath687 the sum converges absolutely whenever @xmath688 because of the estimate in lemma [ lem1 ] and the fact that the distance between the points @xmath190 and @xmath689 grows like @xmath690 as @xmath691 . combining the estimates in lemma [ lemc1 ] with lemma [ lem3 ] we obtain [ greenestcyl ] for any compact subset @xmath692 we have on @xmath693 the following estimates @xmath694 where @xmath695 is the distance between @xmath168 and @xmath696 and @xmath697 . similarly lemma [ lemc1 ] and lemma [ lemc2 ] imply the following estimates [ greenestcylder ] let @xmath698 be a truncated hyperbolic cylinder and let @xmath414 . then for all points @xmath699 and @xmath700 with @xmath701 the following estimates hold . @xmath702 suppose now that @xmath692 is a subset of the hyperbolic cylinder @xmath678 whose boundary @xmath703 is a finite union of geodesic segments . if @xmath704 is a function on @xmath61 , smooth up the the boundary , such that @xmath705 in the interior of @xmath61 . let @xmath706 and @xmath707 be the restriction of @xmath704 to the boundary and its outward normal derivative respectively . of course , as functions on @xmath678 we have @xmath708 in the sense of distributions , where @xmath157 if the distribution defined as @xmath709 from this it follows that @xmath704 in the interior of @xmath61 can be expressed as @xmath710 this is sometimes written in integral form @xmath711 using , lemma [ greenestcyl ] and applying the generalized young inequality we obtain the following estimate . let @xmath713 be the heaviside step function defined by @xmath714 suppose that @xmath313 is a compact connected oriented hyperbolic surface and let @xmath715 be an orthonormal basis of eigenfunctions for the laplace operator on @xmath313 . let @xmath716 be the corresponding sequence of eigenvalues . let @xmath717 be integral kernel of the operator @xmath718 and denote by @xmath719 be its restriction to the diagonal . of course , by the spectral theorem , @xmath720 integrating this over @xmath313 yields the eigenvalue counting function @xmath721 : @xmath722 by construction @xmath723 and @xmath266 are non - decreasing odd functions on @xmath724 . let @xmath725 be twice the injectivity radius at the point @xmath726 . if @xmath170 is the length of a systole we have the estimate @xmath727 . using selberg s pretrace trace formula and finite propagation speed one can easily see that the cosine transform of @xmath728 coincides on the interval @xmath729 with the cosine transform of the function @xmath730 , where @xmath731 is defined by @xmath732 note that since @xmath733 and @xmath734 is non - positive on @xmath735 we obtain the estimate @xmath736 where @xmath737 . let @xmath738 be the first dirichlet eigenvalue of the operator @xmath739 on the interval @xmath740 $ ] and let @xmath741 be the corresponding normalized eigenfunction . we think of @xmath32 as a function on @xmath724 by extending by zero . it can be easily worked out that @xmath738 is the smallest non - zero solution to the equation @xmath742 for @xmath1 . its numerical value is @xmath743 . the estimate @xmath744 is an immediate consequence if the intermediate value theorem . as a test function in the fourier tauberian theorem we use the function @xmath745 . let @xmath746 and @xmath747 be defined as in @xcite as @xmath748 where @xmath749 using @xmath750 and @xmath751 we obtain @xmath752 the fourier tauberian theorem 1.3 in @xcite together with the estimates equations ( 2.9 ) and ( 2.10 ) in @xcite and our estimates ( [ esteem1 ] ) and ( [ esteem2 ] ) imply the following bounds of @xmath266 . setting @xmath763 one obtains a bound for the local heat kernel . the heat trace @xmath764 is obtained by integrating the local heat trace . lemma [ heattrb ] therefore yields bounds on the heat trace in case @xmath763 and on @xmath765 in general . similarly , one gets a bound on the derivative of the eigenfunctions in the following way . fix a point @xmath726 and a unit vector @xmath104 . as above we can define @xmath770 using the pre - trace formula and finite propagation speed it can easily worked out that cosine transform of @xmath771 coincides on the interval @xmath729 with the cosine transform of the function @xmath772 , where @xmath773 is defined by @xmath774 using @xmath775 we obtain in the same way as before @xmath776 where @xmath777 . * acknowledgements . * we would like to thank peter buser for very interesting discussions and also for helping us to compute the fenchel nielsen coordinates of surfaces given in various parametrizations . we are very grateful to ralf aurich for providing the eigenvalues of surfaces for testing purposes . our thanks also go to alex barnett , frank steiner and andreas strmbergsson for useful comments and stimulating discussions . d. a. hejhal , _ on eigenfunctions of the laplacian for hecke triangle groups _ , emerging applications of number theory ( d. hejhal , j. friedman , m. gutzwiller , and a. odlyzko , eds . ) , springer - verlag , 1999 , pp . 291315 . c. klein , a. kokotov , and d. korotkin , _ extremal properties of the determinant of the laplacian in the bergman metric on the moduli space of genus two riemann surfaces _ , math . z. * 261 * ( 2009 ) , no . 1 , 73108 . j. marklof , _ selberg s trace formula : an introduction _ , arxiv : math/0407288 . in hyperbolic geometry and applications in quantum chaos and cosmology , eds . j. bolte and f. steiner , cambridge university press 2011 , pp .	we present a rigorous scheme that makes it possible to compute eigenvalues of the laplace operator on hyperbolic surfaces within a given precision . the method is based on an adaptation of the method of particular solutions to the case of locally symmetric spaces and on explicit estimates for the approximation of eigenfunctions on hyperbolic surfaces by certain basis functions . it can be applied to check whether or not there is an eigenvalue in an @xmath0-neighborhood of a given number @xmath1 . this makes it possible to find all the eigenvalues in a specified interval , up to a given precision with rigorous error estimates . the method converges exponentially fast with the number of basis functions used . combining the knowledge of the eigenvalues with the selberg trace formula we are able to compute values and derivatives of the spectral zeta function again with error bounds . as an example we calculate the spectral determinant and the casimir energy of the bolza surface and other surfaces
for a knot @xmath4a smooth and oriented embedding of the unit circle @xmath5 into the unit @xmath6-sphere @xmath7the _ slice genus _ @xmath8 is the minimal genus of smooth , oriented surfaces @xmath9 in the closed unit @xmath1-ball @xmath10 with oriented boundary @xmath11 . for torus knots , the slice genus is equal to their genus @xmath12 ; i.e. , for coprime positive integers @xmath13 and @xmath14 , one has @xmath15 more generally , the slice genus is known for knots @xmath4 that arise as the transversal intersection of @xmath7 with a smooth algebraic curve @xmath16 in @xmath0 , by kronheimer and mrowka s resolution of the thom conjecture ( * ? ? ? * corollary 1.3 ) : the surface in @xmath17 given as the intersection of @xmath10 with @xmath16 has genus @xmath8 ; see also rudolph s slice - bennequin inequality @xcite . we determine the slice genus for connected sums of torus knots of small braid index ; see corollary [ cor : cobdist ] . the _ cobordism distance _ @xmath18 between two knots @xmath4 and @xmath19 is the minimal genus of smoothly embedded and oriented surfaces @xmath20 in @xmath21 $ ] with boundary @xmath22 such that the induced orientation agrees with the orientation of @xmath19 and disagrees with the orientation of @xmath4 . such a @xmath20 is called a _ cobordism _ between @xmath4 and @xmath19 . equivalently , @xmath18 can be defined as the slice genus of the connected sum @xmath23 of @xmath4 and @xmath24the mirror image of @xmath19 with reversed orientation . cobordism distance satisfies the triangle inequality ; in particular , @xmath25 for all knots @xmath4 and @xmath19 . we call a cobordism between two knots _ algebraic _ if it arises as the intersection of a smooth algebraic curve in @xmath0 with @xmath26 $ ] , where @xmath27 are the @xmath1-balls centered at the origin of radius @xmath28 for some @xmath29 . by the thom conjecture , algebraic cobordisms between two knots @xmath4 and @xmath19 have genus @xmath30 . in particular , the existence of an algebraic cobordism between @xmath4 and @xmath19 does determine their cobordism distance to be @xmath30 . we call a cobordism @xmath20 between two knots @xmath4 and @xmath19 _ optimal _ if its genus @xmath31 is @xmath30 . we address the existence of algebraic and optimal cobordisms for torus knots . [ thm:32 ] for positive torus knots @xmath32 and @xmath33 of braid index 2 and 3 , respectively , the following are equivalent . a. [ 1 ] there exists an optimal cobordism between @xmath32 and @xmath33 ; i.e. @xmath34 b. [ 2]@xmath35 c. [ 3]there exists an algebraic cobordism between @xmath32 and @xmath33 . [ thm:42 ] for positive torus knots @xmath32 and @xmath36 of braid index 2 and 4 , respectively , the following are equivalent . a. there exists an optimal cobordism between @xmath32 and @xmath36 ; i.e. @xmath37 b. @xmath38 c. there exists an algebraic cobordism between @xmath32 and @xmath36 . theorem [ thm:32 ] and theorem [ thm:42 ] are established using the following strategy . if holds , we provide an explicit construction of the optimal cobordism using positive braids . in fact , the optimal cobordisms we find , can be seen in @xmath7 as a sequence of positive destabilizations on the fiber surfaces of the larger - genus knot ; see remark [ rem : posdestab ] . if does not hold , we use the @xmath39-invariant to obstruct the existence of an optimal cobordism . here , @xmath40 is one of a family @xmath41 of concordance invariants introduced by ozsvth , stipsicz , and szab @xcite , which generalize ozsvth and szab s @xmath42-invariant as introduced in @xcite . finally , all optimal cobordisms we construct turn out to be algebraic . we establish a more general result for all knots that arise as the transversal intersection of @xmath7 with a smooth algebraic curve in @xmath0 : the natural way of constructing optimal cobordisms always yields algebraic cobordisms ; see lemma [ lemma : algrealization ] . this brings us to ask : if there exists an optimal cobordism between two positive torus knots , does there exist an algebraic cobordism between them ? the proof of lemma [ lemma : algrealization ] uses realization results of orevkov and rudolph . using deplumbing , we also construct algebraic cobordisms between @xmath32 and @xmath43 ; see section [ subsec : ddto2n ] . this is related to work of orevkov @xcite , see remark [ rem : tmmtot2n ] , and motivated by algebraic geometry questions ; see section [ sec : motivation ] . we now turn to the cobordism distance between torus knots . by gluing together the optimal cobordisms given in theorems [ thm:32 ] and [ thm:42 ] , we will obtain the following . [ cor : cobdist ] let @xmath4 and @xmath19 be torus knots such that the sum of their braid indices is 6 or less . then we have the following formula for their cobordism distance . @xmath44 the values of the concordance invariants @xmath42 and @xmath39 that arise in corollary [ cor : cobdist ] are explicitly calculable : for positive torus knots ( for their negative counterparts , which are obtained by taking the mirror image , the sign changes ) , one has @xmath45{ozsvathszabo_03_kfhandthefourballgenus}},\ ] ] for coprime positive integers @xmath13 and @xmath14 , and @xmath46 for positive integers @xmath47 . the @xmath39-values for torus knots of braid index 2 follow , for example , from the fact that @xmath39 equals half the signature for alternating knots ( * ? ? ? * theorem 1.14 ) . the other @xmath39-values in can be derived from the inductive formula for @xmath39 provided by ozsvth , stipsticz and szab ( * ? ? ? * theorem 1.15 ) . we do this in section [ sec : upsilonfortorusknots ] . of course , part of the statement of corollary [ cor : cobdist ] was known before ; for example in the cases covered by the following remark . [ rem : cor : cobdist ] if @xmath4 and @xmath19 are positive and negative torus knots , respectively , one has @xmath48 by the thom conjecture . if @xmath4 and @xmath19 are both positive ( both negative ) torus knots of the same braid index , then there exists an optimal cobordism between them ; compare example [ ex : incomptnm < tabifn < am < b ] . determining the cobordism distance between all torus knots seems out of reach . however , a coarse estimation of the cobordism distance between torus knots was provided by baader @xcite . the study of optimal and algebraic cobordisms between torus knots seems natural from a knot theoretic point of view . we discuss additional motivation from algebraic geometry questions in section [ sec : motivation ] . section [ sec : notations ] recalls the notions of positive and quasi - positive braids . in section [ sec : algrealization ] we show that optimal cobordisms given by quasi - positive braid sequences are algebraic . in section [ sec : constructionofcobs ] optimal cobordisms are constructed and used to prove theorem [ thm:32 ] , theorem [ thm:42 ] , and corollary [ cor : cobdist ] . section [ sec : upsilonfortorusknots ] provides the @xmath3-values for torus knots of braid index @xmath6 and @xmath1 . * acknowledgements : * i thank sebastian baader for sharing his insight into cobordisms and for his ongoing support . thanks also to adam levine for pointing me towards the @xmath3-invariant , and immanuel stampfli and andrew yarmola for valuable suggestions . finally , i wish to thank the referee for helpful suggestions and corrections . in this section , we discuss motivations to study algebraic and optimal cobordisms between torus knots coming from singularity theory . mathematically , the rest of the paper is independent of this . we consider isolated singularities on algebraic curves in @xmath0 and we denote singularities by the function germs that define them . a general question asks , what ( topological ) type of singularities can occur on an algebraic curve @xmath16 in @xmath0 given as the zero - set of a square - free polynomial @xmath50 in @xmath51 $ ] of some fixed degree @xmath52 ; see e.g. greuel , lossen , and shustin s work @xcite . even for simple singularitiesthose corresponding to dynkin diagrams @xcitea lot is unknown . for the @xmath53-singularitiesthe simple singularities given by @xmath54the following bounds were provided by gusen - zade and nekhoroshev @xmath55 where @xmath56 denotes the maximal integer @xmath47 such that @xmath53 occurs on a degree @xmath52 curve . in fact , orevkov improved the lower bound to @xmath57 recall that , for a singularity at @xmath58 , its _ link of singularity _ is the link obtained as the transversal intersection of @xmath16 with the small @xmath6-sphere @xmath59 for small enough @xmath60 ; see milnor @xcite . similarly , the _ link at infinity _ of an algebraic curve @xmath16 is defined to be the transversal intersection of @xmath16 with the @xmath6-sphere @xmath61 of radius @xmath62 for @xmath62 large enough ; see e.g. neumann and rudolph @xcite . prototypical examples of plane curve singularities are the singularities @xmath63 , where @xmath13 and @xmath14 are positive integers . they have the torus link @xmath64 as link of singularity . up to topological type , singularities are determined by their link of singularity and the links that arise as links of singularities are fully understood : their components are positive torus knots or special cables thereof ; see e.g. brieskorn and knrrer s book @xcite . in the algebraic setting it is natural to consider links not only knots . algebraic cobordisms between links are defined as for knots . optimal cobordisms are defined via euler characteristic instead of the genus ; see section [ sec : algrealization ] . [ obs : algcob ] the existence of a singularity on a curve of degree @xmath52 implies the existence of an optimal cobordism from the link of the singularity to @xmath65 . in particular , there exists an optimal cobordism from @xmath66 to @xmath65 . observation [ obs : algcob ] motivated our study of optimal cobordisms from @xmath32 to @xmath65 and @xmath67 . in section [ subsec : ddto2n ] we show that there exist algebraic cobordisms between @xmath32 and @xmath65 ( @xmath67 ) if @xmath68 . in particular , no obvious topological obstruction exists to having @xmath69 ; compare also @xcite . observation [ obs : algcob ] allows to give a knot theoretic proof of the upper bound in ; see remark [ rem : subwordsoftmm ] . to establish observation [ obs : algcob ] , we note that , whenever a singularity occurs on an algebraic curve @xmath16 , we get an algebraic cobordism from the link of the singularity @xmath4 to the link at infinity @xmath19 of @xmath16 . for this , choose a small sphere @xmath70 and a large sphere @xmath71 that intersect @xmath16 transversally in @xmath4 and @xmath19 , respectively . let @xmath72 be another algebraic curve . by transversality , @xmath16 and @xmath72 intersects @xmath73 and @xmath71 in the same links as long as @xmath12 and @xmath50 are `` close '' . to be precise , this is certainly true if @xmath74 and @xmath75 is small . for generic @xmath76 , @xmath72 is smooth ; thus , @xmath72 yields an algebraic cobordism between @xmath4 and @xmath77 . furthermore , there is an algebraic cobordism from @xmath19 to @xmath65 . this follows by using that the link at infinity of @xmath78 is @xmath65 , for generic @xmath79 , while @xmath80 is @xmath19 for @xmath81 small enough ; and then arguing as above . gluing the two algebraic cobordisms together yields an optimal cobordism from @xmath4 to @xmath65 . a related question asks about the existence of adjacencies between singularities . fixing a singularity @xmath50 , another singularity @xmath12 is said to be _ adjacent _ to @xmath50 if there exists a smooth family of germs @xmath82 such that @xmath83 and @xmath84 for small enough non - zero @xmath76 . there are different notions of equivalence @xmath85 yielding different notions of adjacencies . however , as long as @xmath12 defines a simple singularity the notions all agree . see siersma s work for a discussion of these notions @xcite and compare also with arnol@xmath86d s work , who was the first to fully describe adjacency between simple singularities ( * ? ? ? * corollary 8.7 ) . a modern introduction to singularities and their deformations is provided by greuel , lossen and shustin @xcite . if @xmath12 is adjacent to @xmath50 , then there exists an algebraic cobordism between their links of the singularity ( given by @xmath87 for @xmath76 and @xmath88 small as a similar argument as above shows ) . the adjacency question is mostly unresolved if @xmath50 is not a simple singularity . a natural first case to consider is @xmath89 for fixed @xmath90 and to ask : given a positive integer @xmath14 , which @xmath91-singularities are adjacent to @xmath92 ? theorem [ thm:32 ] and theorem [ thm:42 ] can be seen as answering analogs of this question for @xmath13 equal to @xmath6 and @xmath1 , respectively . to set notions , we shortly recall artin s braid group @xcite ; a nice reference for braids is birman s book @xcite . let us fixe a positive integer @xmath93 . the standard group presentation for the _ braid group on @xmath93 strands _ , denoted by @xmath94 , is given by generators @xmath95 subject to the _ braid relations _ @xmath96 elements @xmath97 of @xmath94 , called _ braids _ or _ n - braids _ , have a well - defined ( algebraic ) length @xmath98 , given by the number of generators minus the number of inverses of generators in a word representing @xmath97 . more geometrically , a @xmath93-braid @xmath97 can be viewed as an isotopy class of an oriented compact @xmath99-submanifold of @xmath100\times \c$ ] such that the projection to @xmath100 $ ] is a @xmath93-fold orientation - preserving regular map and @xmath97 intersects @xmath101 and @xmath102 in @xmath103 and @xmath104 , respectively , where @xmath105 is a subset of @xmath49 consisting of @xmath93 complex numbers with pairwise different real part . the above standard generators @xmath106 are identified with the braid that exchanges the @xmath107th and @xmath108th ( with respect to order induced by the real order ) point of @xmath105 by a half - twist parameterized by @xmath100 $ ] and the group operations is given by stacking braids on top of each other . the _ closure _ @xmath109 of @xmath97 is the closed @xmath99-submanifold in @xmath110 obtained by gluing the top of @xmath111\times\c$ ] to its bottom . a closed braid @xmath109 yields a link in @xmath7 , also denoted by @xmath109 , via a fixed standard embedding of the solid torus @xmath110 in @xmath7 . the _ braid index _ of a link is the minimal number @xmath93 such that @xmath77 arises as the closure of a @xmath93-braid . _ positive braids _ are the elements of the semi - subgroup @xmath112 that is generated by all the generators @xmath106 . _ positive torus links _ are examples of links that arise as closures of positive braids : for positive integers @xmath13 and @xmath14 , the closure of @xmath113 is @xmath64 , which is a knot , called a _ positive torus knot _ , if and only if @xmath13 and @xmath14 are coprime . the braid index of @xmath64 is @xmath114 . rudolph introduced _ _ quasi - positive braids__the elements of the semi - subgroup of @xmath94 generated by all conjugates of the generators @xmath106 ; i.e. the braids given by _ quasi - positive braid words _ @xmath115 ; compare @xcite . a knot or link is called _ quasi - positive _ if it arises as the closure of a quasi - positive braid . a quasi - positive braid @xmath97 has an associated canonical ribbon surface @xmath116 embedded in @xmath10 with the closure of @xmath97 as boundary , which can be seen in @xmath7 given by @xmath93 discs , one for every strand , and @xmath98 ribbon bands between the discs . in particular , the euler characteristic @xmath117 of @xmath116 is @xmath118 . by the slice - bennequin inequality @xcite , @xmath119 equals @xmath120the maximal euler characteristic among all oriented and smoothly embedded surfaces @xmath9 ( without closed components ) in @xmath10 such that @xmath121 is the link @xmath109 . rudolph established that all quasi - positive links arise as the transversal intersection of a smooth algebraic curve in @xmath0 with @xmath7 @xcite . boileau and orevkov proved that this is a characterization of quasi - positive links ( * ? ? ? * theorem 1 ) . this section is concerned with establishing the following realization lemma . [ lemma : algrealization ] let @xmath122 and @xmath123 be quasi - positive @xmath93-braid words . if @xmath123 can be obtained from @xmath122 by applying a finite number of braid group relations , conjugations , and additions of a conjugate of a generator anywhere in the braid word ; then there exists an algebraic cobordism @xmath20 between the links obtained as the closures of the @xmath124 . in fact , @xmath20 is given as the zero - set of a polynomial in @xmath51 $ ] of the form @xmath125 where the @xmath126 are polynomials . let @xmath122 and @xmath123 be quasi - positive braid words given as in lemma [ lemma : algrealization ] ; i.e. there is a sequence of @xmath93-braid words @xmath127 starting with @xmath122 ending with @xmath123 such that @xmath128 and @xmath129 either define the same conjugacy class in @xmath94 or @xmath129 is obtained by adding a generator @xmath106 somewhere in @xmath128 . there is an associated cobordism @xmath20 between @xmath130 and @xmath131 given ( as a handle decomposition ) by 1-handle attachments corresponding to every generator that is added . the cobordism @xmath20 is _ optimal _ ; i.e. it has euler characteristic @xmath132 and does not have closed components ( this is the sensible extension of the notion of optimal cobordisms to links ) . in fact , although not made explicit , the proof of lemma [ lemma : algrealization ] given below does show that this @xmath20 is algebraic . all optimal cobordisms we construct in section [ sec : constructionofcobs ] arise as described above . we see lemma [ lemma : algrealization ] as evidence that the following question might have a positive answer . [ q : algrealization ] are the two necessary conditions for the existence of an algebraic cobordism between two knotsthe knots are quasi - positive and there exists an optimal cobordism between themsufficient ? the proof of lemma [ lemma : algrealization ] occupies the rest of this section and uses rudolph diagrams . only the statement of lemma [ lemma : algrealization ] is used in the rest of the paper . to set notation and for the reader s convenience , we recall the notion of rudolph diagrams , following @xcite and @xcite . for a square - free algebraic function @xmath133 of the form @xmath134,\ ] ] we study the following subsets of @xmath49 . the finite set @xmath135 of all @xmath136 such that some of the @xmath93 solutions @xmath137 of @xmath138 coincide . the semi real - analytic set @xmath139 of all @xmath136 such that the @xmath93 solutions of @xmath138 are all different , but do not have @xmath93 distinct real parts . their union @xmath140 is denoted by @xmath141 . [ ex : g(f ) ] if @xmath50 is @xmath142 , then @xmath143 , @xmath144 , and @xmath145 . let @xmath146 denote the zero - set of @xmath50 . for an oriented simple closed curve @xmath147 in @xmath148 , the intersection @xmath149 is a closed @xmath93-braid via the identification @xmath150 . similarly , for every oriented arc @xmath151 in @xmath152 with endpoints in @xmath153 ( which guarantees that at endpoints the @xmath93-solutions have different real parts ) , the intersection @xmath154 is a @xmath93-braid by identifying @xmath155 with @xmath100\times \c$ ] . note that for this to be well - defined , the identification should preserve the order of the real parts in the second factor . an endpoint - fixing isotopy of two arcs and an isotopy of two simple closed curves in @xmath156 correspond to an isotopy of braids and an isotopy of closed braids , respectively . any choice of convention not made explicit so far is chosen such that in example [ ex : g(f ) ] the oriented arc starting at @xmath157 and ending at @xmath158 yields the @xmath159-braid @xmath160 . let @xmath161 be the projection to the first coordinate . we will only consider @xmath50 such that @xmath162 defines a non - singular algebraic curve @xmath163 and such that for every @xmath136 in @xmath135 the intersection @xmath164 consists of exactly @xmath165 points ; i.e. fixing an @xmath136 in @xmath135 gives a polynomial in @xmath166 with precisely one repeated root of multiplicity two . rudolph observed that @xmath167 naturally carries the structure of an oriented , @xmath168-labeled graph that describes @xmath16 up to @xmath169-preserving smooth isotopy in @xmath0 . after a generic small linear coordinate change ( to rule out pathologies ) , the vertices are locally given as in figure [ fig : vertices ] . the elements of @xmath135 are the @xmath99-valent vertices . the edges are the connected components of the semi real - analytic open subset of @xmath139 given by those @xmath136 that have precisely @xmath165 different real parts among the real parts of the solutions @xmath170 to @xmath138 . some of the edges tend to infinity instead of ending at a vertex . an edge @xmath171 is labeled as follows . for @xmath136 in @xmath172 , the @xmath93 solutions @xmath173 of @xmath174 can be indexed such that the index order agrees with the order given by their real parts ; i.e. @xmath175 for some @xmath176 in @xmath168 . the edge @xmath171 is labeled with @xmath176 . the edges are oriented as follows . pick a small oriented arc @xmath177 that meets @xmath171 transversally in a point @xmath136 . the braid associated with @xmath151 is either @xmath178 or @xmath179 ( the @xmath176th and @xmath180th solution exchange their real - part order while passing through @xmath136 ) . orient @xmath171 such that , if the orientation of @xmath171 followed by the orientation of @xmath147 gives the complex orientation of @xmath49 , then the braid corresponding to the transverse arc is @xmath181 ( rather than for @xmath182 ) . in particular , edges are oriented to point away from the @xmath99-valent vertices . the oriented , labeled graph @xmath141 describes @xmath16 up to @xmath169-preserving isotopy ; in particular , it describes all closed braids given by intersecting @xmath16 with cylinders . for a fixed embedded curve in @xmath147 in @xmath152 with a marked start point @xmath183 , one gets an explicit procedure , how to read off a braid word for the braid @xmath97 corresponding to the arc starting and ending at @xmath13 going counter - clockwise around @xmath147 : by a small isotopy of @xmath147 in @xmath152 , we may assume that @xmath147 meets @xmath141 transversally and in edges only . starting at @xmath13 we move counter - clockwise around @xmath147 . whenever we cross an edge @xmath171 transversally at a point @xmath136 , we write down the generator @xmath178 or its inverse @xmath179 depending on whether the orientation at @xmath136 given by @xmath141 and @xmath147 agrees or disagrees with the complex orientation of @xmath49 . the study of the graphs @xmath141 motivates the following definition . fix some surface @xmath184 with boundary . in fact , we will only consider cases where @xmath184 is either @xmath185 a _ rudolph diagram _ on @xmath184 is an oriented , @xmath168-labeled graph @xmath186 with smooth edges ( we also allow smooth closed cycles ) that enters and exits the boundary of @xmath184 transversely and is locally modelled on graphs @xmath141 coming from an algebraic function @xmath187 $ ] as above ; i.e. locally around a vertex @xmath186 is given as in figure [ fig : vertices ] . we denote the set of @xmath99-valent vertices by @xmath188 . of course , a huge source of examples are obtained by embedding ( or immersing ) @xmath184 in @xmath49 and using this embedding to define @xmath186 as the pull back of @xmath141 for some algebraic function @xmath50 . any closed curve @xmath147 missing @xmath99-valent vertices defines a closed braid @xmath109 by isotoping @xmath147 to meet @xmath186 transversally in edges and then reading off a braid word @xmath97 for that closure as described in the @xmath141-case . a rudolph diagram is said to be _ smooth _ if it contains only @xmath99-valent vertices . [ ex : realizingqpasrd ] given a quasi - positive braid word @xmath189 , there exists a smooth rudolph diagram on @xmath190 such that braid word read off when following @xmath191 is @xmath97 . figure [ fig : conjugaterealization ] illustrates how one factor @xmath192 is realized . orevkov describes which smooth rudolph diagrams on @xmath190 arise as @xmath141 . [ prop : rudolphrealization ] ( * ? ? ? * proposition 2.1 ) let @xmath186 be a smooth rudolph diagram on @xmath190 . there exists an algebraic function @xmath193\ ] ] such that @xmath186 is isotopic to @xmath194 if and only if @xmath195 contains no closed cycles . as pointed out by orevkov , rudolph ( implicitly ) proved such a statement while establishing the main theorem of @xcite . [ rem : cycleremove ] given a smooth rudolph diagram @xmath186 on @xmath190 one can remove all closed cycles without changing the closed braids associated with closed curves in @xmath190 . for a rudolph diagram @xmath186 on @xmath196 , we denote by @xmath122 and @xmath123 the two braids defined by @xmath186 via reading off braid words following the inner and outer boundary of @xmath196 counter - clockwise starting at @xmath99 and @xmath159 , respectively . for @xmath122 and @xmath123 to be well - defined , we ask that @xmath186 does not meet @xmath99 or @xmath159 , which from now on is imposed on every rudolph diagram . for the proof of lemma [ lemma : algrealization ] , we need the following . if a braid @xmath97 is obtained from a braid @xmath151 as described in lemma [ lemma : algrealization ] , then there exists a rudolph diagram on @xmath196 such that @xmath197 and @xmath198 . the rest of this subsection provides one way of making this statement precise . [ rem : genericoutwarddiagram ] any rudolph diagram @xmath186 on @xmath190 or @xmath196 can be isotoped such that it is _ outward - oriented _ , which is defined as follows : all but a finite number of circles around the origin intersect @xmath186 transversally in edges . furthermore , the finite exceptional circle meet @xmath186 transversally in edges except in one point @xmath136 , which falls in one of two categories . either @xmath136 lies in the interior of an edge and the radial function restricted to that edge has a strict local extremum . or @xmath136 is a vertex and locally around @xmath136 the rudolph diagram @xmath186 and the exceptional circle behave as described in figure [ fig : verticescc ] . finally , the positive real ray @xmath199 meets @xmath186 transversally in edges and away from the finite number of exceptional circles . locally around points in @xmath200 the radial function increases on @xmath186 when following the orientation . an example is provided in figure [ fig : genericoutward ] . given an outward - oriented rudolph diagram @xmath186 on @xmath196 , let @xmath201 denote the radii corresponding to exceptional circles or points where @xmath186 meets @xmath199 . for @xmath81 in @xmath202\setminus\{r_1,\cdots r_k\}$ ] , we denote by @xmath203 the braid read off when following the counter - clockwise oriented circle of radius @xmath81 with @xmath81 as marked starting point . we associate to @xmath186 the following finite sequence of braid words @xmath204 where @xmath205 in particular , the sequence starts and ends with @xmath122 and @xmath123 , respectively . [ obs : consecbraids ] for all @xmath206 , the braid word @xmath207 is obtained from @xmath208 by one of the following operations , for some @xmath209 with @xmath210 : a. [ i ] adding or removing subwords @xmath211 or @xmath212 ; b. [ ii ] performing one braid relation ; i.e. replacing @xmath213 by @xmath214 or replacing @xmath215 by @xmath216 ; c. [ iii ] changing a braid word of the form @xmath217 to @xmath218 or vice versa ; d. [ iv ] adding @xmath106 somewhere in the braid word . [ rem : consecbraids ] note that two braid words can be connected with a sequence using and if and only if they define the same braid . two braid words can be connected with a sequence using , , and if and only if they define the same closed braid . and two braid words can be connected with a sequence using through if and only if they are connected as described in lemma [ lemma : algrealization ] . conversely , a sequence of braid words as described in observation [ obs : consecbraids ] yields a rudolph diagram on @xmath196 . this amounts to the following : [ prop : assobraidseq ] the assignment given by yields an one - to - one correspondence between outward - oriented rudolph diagrams on @xmath196 , up to isotopy through outward - oriented rudolph diagrams , and finite non - empty sequences of braid words @xmath219 such that @xmath220 is obtained from @xmath221 by one of the operations , , and described in observation [ obs : consecbraids ] . after this translation of sequences of braid words to rudolph diagrams we need a final ingredient to prove lemma [ lemma : algrealization ] : [ prop : smoothingofrd ] let @xmath186 be a rudolph diagram on @xmath196 . there exists a smooth rudolph diagram @xmath222 on @xmath196 satisfying the following : @xmath186 and @xmath222 are identical in a neighborhood of the inner boundary @xmath5 . in particular , the braid words @xmath122 and @xmath223 corresponding to @xmath5 are the same . the braids @xmath123 and @xmath224 corresponding to the outer boundary @xmath225 have the same closure . let @xmath226 be the set of @xmath99-valent vertices in @xmath186 . for every @xmath227 in @xmath135 , we choose an embedded arc @xmath228 in @xmath196 that connects @xmath227 to the inner boundary @xmath5 of @xmath196 such that @xmath228 intersects @xmath186 in the interior of @xmath196 transversally and outside of vertices ( except at @xmath227 of course ) . furthermore , we arrange that all the arcs @xmath229 are pairwise disjoint . a neighborhood @xmath230 in @xmath196 of the union @xmath231 defines an annulus on which @xmath186 is smooth . the boundary of @xmath230 has two components : @xmath5 and a curve that is isotopic to @xmath225 in @xmath232 . therefore , we obtain a rudolph diagram on @xmath196 as wanted by identifying @xmath230 with @xmath196 via a diffeomorphism that is the identity in a neighborhood of @xmath5 . this is illustrated in figure [ fig : smoothingrd ] . let @xmath122 and @xmath123 be quasi - positive braid words satisfying the assumptions in lemma [ lemma : algrealization ] . we find a finite sequence of braid words as described in observation [ obs : consecbraids ] that starts with @xmath122 and ends with @xmath123 . let @xmath186 be the corresponding rudolph diagram on @xmath196 provided by proposition [ prop : assobraidseq ] . by proposition [ prop : smoothingofrd ] we may assume that @xmath186 is smooth ( this may change @xmath123 but the corresponding closed braid remains the same ) . since @xmath122 is quasi - positive , there is a smooth rudolph diagram @xmath222 on @xmath190 such that @xmath122 is the braid word read off when following the boundary of @xmath190 by example [ ex : realizingqpasrd ] . we glue @xmath233 and @xmath234 together along @xmath5 to get ( by rescaling ) a smooth rudolph diagram @xmath62 on the disk @xmath190 ; see left - hand side of figure [ fig : uniformization ] . next , we remove all closed cycles in @xmath62 . this might change the braid word @xmath122 but not the closed braid it represents by remark [ rem : cycleremove ] . by proposition [ prop : rudolphrealization ] , there exists an algebraic function @xmath50 such that @xmath235 after an isotopy of @xmath62 . the latter isotopy yields an embedding of @xmath196 in @xmath49 ( we denote its image again by @xmath196 ) such that @xmath236 . by the uniformization theorem for open annuli ( see e.g. ( * ? ? ? * theorem 10 ) ) , there exists a biholomorphic map @xmath237 from the interior of @xmath238 to an open annulus @xmath239 with concentric boundary circles centered at the origin . setting @xmath240 defines holomorphic map on @xmath241 of the form @xmath242 we choose two concentric circles @xmath243 and @xmath244 in @xmath245 such that their preimages under @xmath237 are curves @xmath246 and @xmath247 in @xmath196 for which the closed braids @xmath248 are @xmath249 . this is , for example , achieved by choosing @xmath250 and @xmath251 close to the radii of the inner and the outer boundary of @xmath245 , respectively ; see figure [ fig : uniformization ] . therefore , @xmath252 intersects the cylinders @xmath253 transversely in closed braids and those closed braids are @xmath249 since @xmath254 is the image of @xmath248 under the biholomorphic map @xmath255 by polynomial approximation of the holomorphic coefficients @xmath256 of @xmath257 , we find a polynomial @xmath258 with @xmath259 $ ] such that its zero - set @xmath72 intersects the above cylinders transversally in the same closed braids as @xmath252 . we replace the cylinders @xmath260 with cylinders @xmath261 , which for large enough @xmath62 intersect @xmath72 in the same closed braids as the @xmath260 . finally , we set @xmath262 and conclude the proof by noticing that the @xmath6-spheres @xmath263 of radius @xmath28 intersect the zero - set @xmath264 in the links that are the closures of @xmath124 since rescaling the @xmath166-coordinate by the factor @xmath265 maps @xmath266 onto @xmath267 . in this section , we construct several families of optimal cobordisms between torus knots , which are also algebraic by lemma [ lemma : algrealization ] . it came as a surprise to the author that ozsvth , stipsicz , and szab s @xmath3-invariant shows that the constructions for torus knots of braid index 4 or less can not be improved . for links @xmath4 and @xmath19 that are closures of positive braids , we say @xmath4 is _ subword - adjacent _ to @xmath19 , denoted by @xmath268 , if there are positive @xmath93-braids @xmath122 and @xmath123 , for some integer @xmath93 , such that @xmath122 can be obtained from @xmath123 by successively deleting generators . here , deleting a generator in a positive braid @xmath97 means removing an @xmath106 in a positive braid word that represents @xmath97 . we think of subword - adjacency as a combinatorial toy model for adjacency of singularities ( as described in section [ sec : motivation ] ) , hence the name . [ rem : subwordadjacency ] if a positive @xmath93-braid @xmath122 is obtained from a positive @xmath93-braid @xmath123 by deleting positive generators , then @xmath123 can be obtained from @xmath122 as described in lemma [ lemma : algrealization ] . therefore , if @xmath4 is subword - adjacent to @xmath19 , then there exists an algebraic cobordism between them , by lemma [ lemma : algrealization ] . [ rem : posdestab ] in what follows we consider positive braids @xmath97 with non - split closure ; in particular , their closures @xmath109 are fibered ; see stallings @xcite . in this case , the optimal cobordism provided by a subword - adjacency can be understood on the fiber surfaces : removing a generator in a positive braid @xmath97 corresponds to deplumbing a positive hopf band on its fiber surface @xmath116 . in other words , if @xmath269 is subword - adjacent to @xmath270 , then the open book of @xmath7 with binding @xmath4 is obtained from the open book of @xmath7 with binding @xmath19 by @xmath271 positive destabilizations . in this section , we use fence diagrams to represent positive braids . i.e. in a braid diagram , positive crossings @xmath272%{\vdots } ! { \vcrossneg } } \;$ ] are replaced with horizontal intervals @xmath272!{\xcapv[1]@(0 ) } [ lu]!{\xcapv[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0)}}\hspace{-0.5pc}$ ] ; see rudolph @xcite . for example , the positive 3-braid @xmath273 is represented by @xmath274!{\xcapv[1.5]@(0 ) } [ lu]!{\xcapv[1.5]@(0 ) } [ lu]!{\xcapv[1.5]@(0 ) } [ u(0.5)r ] ! { \xcaph[1]@(0 ) } [ d(0.5)ll ] ! { \xcaph[1]@(0 ) } % [ lll ] ! { \xcaph[1]@(0 ) } % [ lll ] ! { \xcaph[1]@(0 ) } } \hspace{-1pc}$ ] instead of the braid diagram @xmath275%{\vdots } [ d(1.4)]!{\vcrossneg } [ ul]!{\xcapv[1]@(0 ) } [ ld ] [ ur]!{\vcrossneg } [ urr]!{\xcapv[1]@(0 ) } [ ld ] % [ u]!{\vcrossneg } % [ ul]!{\xcapv[1]@(0)}[dr ] % [ urr]!{\xcapv[1]@(0 ) } [ ld ] % { \vdots } } \hspace{-0.5pc}$ ] . simple examples of a subword - adjacency , which yield well - known optimal cobordisms , are the following . [ ex : incomptnm < tabifn < am < b ] let @xmath276 be positive integers . if @xmath277 and @xmath278 , then @xmath279 is subword - adjacent to @xmath280 . this subword - adjacency is obtained by deleting generators in the positive @xmath281-braid word @xmath282 , which has closure @xmath280 , until one reaches a positive braid word with closure @xmath279 . figure [ fig : subwordadjacency ] illustrates this for the adjacency @xmath283 . @xmath284!{\xcapv[9.5]@(0 ) } [ lu]!{\xcapv[9.5]@(0 ) } [ lu]!{\xcapv[9.5]@(0 ) } [ lu]!{\xcapv[9.5]@(0 ) } [ lu]!{\xcapv[9.5]@(0 ) } [ lu]!{\xcapv[9.5]@(0 ) } [ lu]!{\xcapv[9.5]@(0 ) } [ d(2 ) ] ! { \color{red}\xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(3.5)llllll ] ! { \xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(2)lll ] ! { \xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(2)lll ] ! { \xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(2)lll ] ! { \xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(2)lll ] ! { \xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(-0.5)llllll ] ! { \color{black}\xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } % [ u(0.5 ) ] ! { \color{red}\xcaph[1]@(0 ) } % [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } % [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(2)lll ] ! { \color{black}\xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } % [ u(0.5 ) ] ! { \color{red}\xcaph[1]@(0 ) } % [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } % [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(2)lll ] ! { \color{black}\xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } % [ u(0.5 ) ] ! { \color{red}\xcaph[1]@(0 ) } % [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } % [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(2)lll ] ! { \color{black}\xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } % [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } % [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } % [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(2)lll ] ! { \color{black}\xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } } \hspace{-6.2pc } \longrightarrow ~\xygraph { ! { 0;/r0.7pc/ : } [ u(3)]!{\xcapv[6]@(0 ) } [ lu]!{\xcapv[6]@(0 ) } [ lu]!{\xcapv[6]@(0 ) } [ lu]!{\xcapv[6]@(0 ) } [ lu]!{\xcapv[6]@(0 ) } [ lu]!{\xcapv[6]@(0 ) } [ lu]!{\xcapv[6]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(2)lll ] ! { \xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(2)lll ] ! { \xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(2)lll ] ! { \xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(2)lll ] ! { \xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } } \hspace{-3pc},\ ] ] the subword - adjacencies given in example [ ex : incomptnm < tabifn < am < b ] , have analogs in the algebraic adjacency setting since it is easy to write down an adjacency from @xmath285 to @xmath286 . a proposition due to baader provides examples of subword - adjacencies that are more interesting . [ prop : fibersurface : a , bc->b , ac ] ( * ? ? ? * proposition 1 ) let @xmath287 be positive integers with @xmath288 . then @xmath289 is subword - adjacent to @xmath290 . however , again there exists an algebraic adjacency from @xmath291 to @xmath292 ; see ( * ? ? ? * proposition 23 ) , which yields an algebraic cobordisms from @xmath289 to @xmath290 without appealing to lemma [ lemma : algrealization ] . after these first examples , we proceed with subword adjacencies between torus knots that turn out to be optimal and that , to the author s knowledge , are not known to have algebraic adjacency analogs . the following propositions provide all optimal cobordisms that are needed to establish theorem [ thm:32 ] , theorem [ thm:42 ] , and corollary [ cor : cobdist ] . [ prop : subsurfacesofindex3toruslinks ] let @xmath93 and @xmath293 be positive integers . if @xmath294 , then the torus link @xmath32 is subword - adjacent to the torus link @xmath33 . [ prop : t42m > t25 m ] let @xmath93 and @xmath293 be positive integers . if @xmath295 , then the torus link @xmath32 is subword - adjacent to the torus link @xmath36 . it is part of the statement of theorems [ thm:32 ] and [ thm:42 ] that propositions [ prop : subsurfacesofindex3toruslinks ] and [ prop : t42m > t25 m ] can not be improved , at least when the involved links are knots . this is a consequence of the cobordism distance bound @xmath296 provided in ( * ? ? ? * theorem 1.11 ) , generalizing the @xmath42-bound in @xcite . before proving propositions [ prop : subsurfacesofindex3toruslinks ] and [ prop : t42m > t25 m ] , we use them and to prove theorem [ thm:32 ] , theorem [ thm:42 ] , and corollary [ cor : cobdist ] . by remark [ rem : subwordadjacency ] , the fact that implies is an immediate consequence of proposition [ prop : subsurfacesofindex3toruslinks ] and proposition [ prop : t42m > t25 m ] , respectively . by the thom conjecture , implies . therefore , it remains to show that implies . throughout the proof we have @xmath297 , for some odd integer @xmath298 , and @xmath19 is a torus knot of braid index 3 or 4 . we assume towards a contradiction that does not hold and calculate that @xmath299 which contradicts by . we do this according to the 3 cases @xmath300 , @xmath301 ( theorem [ thm:32 ] ) ; and @xmath302 ( theorem [ thm:42 ] ) , where @xmath47 is a positive integer : for @xmath301 , we have that fails precisely when @xmath303 this shows that , if fails , then holds . similarly , for @xmath300 and @xmath302 , we have that fails precisely when @xmath304 as before this shows that , if fails , then holds . for torus knots @xmath4 and @xmath19 such that the sum of their braid indices is @xmath305 or less , we want to establish @xmath306 by , it suffices to find a cobordism @xmath20 between @xmath4 and @xmath19 with genus @xmath307 if @xmath4 and @xmath19 are torus knots of opposite sign , then we have that @xmath308 where the second equality invokes the thom conjecture ; compare remark [ rem : cor : cobdist ] . if @xmath4 and @xmath19 are positive torus knots that have the same braid index , then @xmath309 where the second equality follows from example [ ex : incomptnm < tabifn < am < b ] ; compare remark [ rem : cor : cobdist ] . therefore , if @xmath4 and @xmath19 are torus knots of opposite sign or torus knots with the same braid index , then @xmath310 since @xmath311 or @xmath312 , respectively ; compare ( * ? ? ? * corollary 1.7 ) . thus , after taking mirror images of @xmath4 and @xmath19 , if necessary , we may assume that @xmath4 and @xmath19 are both positive torus knots such that @xmath4 has braid index @xmath159 and @xmath19 has braid index 3 or 4 . let @xmath93 and @xmath47 be the positive integers such that @xmath4 is @xmath32 and @xmath19 is @xmath313 , @xmath314 , or @xmath315 . furthermore , we do not need to consider the cases covered by theorems [ thm:32 ] and [ thm:42 ] ; i.e. when , , and of theorems [ thm:32 ] and [ thm:42 ] , respectively , are satisfied . we first consider @xmath301 . by proposition [ prop : subsurfacesofindex3toruslinks ] there exist a cobordism @xmath316 from @xmath19 to @xmath317 which is optimal ; i.e. @xmath316 has euler characteristic @xmath318 note that @xmath319 since we are assuming that there does not exist an optimal cobordism from @xmath4 to @xmath19 ; compare in theorem [ thm:32 ] . by example [ ex : incomptnm < tabifn < am < b ] , we have a cobordism @xmath320 from @xmath317 to @xmath4 of euler characteristic @xmath321 gluing @xmath316 and @xmath320 together yields a cobordism @xmath20 of euler characteristic @xmath322 between @xmath19 and @xmath4 . thus , @xmath323 similarly , if @xmath300 or @xmath302 , we get a cobordism from @xmath19 to @xmath324 with euler characteristic @xmath325 by proposition [ prop : subsurfacesofindex3toruslinks ] or proposition [ prop : t42m > t25 m ] , respectively , and a cobordism from @xmath324 to @xmath4 with euler characteristic @xmath326 as before , we combine these two cobordisms to a cobordism @xmath20 from @xmath19 to @xmath4 with @xmath327 we now provide proofs of proposition [ prop : subsurfacesofindex3toruslinks ] and proposition [ prop : t42m > t25 m ] . we thank sebastian baader for important inputs for these proofs . we denote the @xmath6-strand full twist @xmath328 by @xmath329 . the full twist commutes with every other @xmath6-braid . let us first consider the case where @xmath330 for some positive integer @xmath331 . the torus link @xmath332 is the closure of @xmath6-braid @xmath333 . note that @xmath334 adding another full twist yields @xmath335 and inductively we get @xmath336 . the subword @xmath337 occurs @xmath338 times in @xmath339 . applying @xmath338 times the braid relation @xmath340 gives the @xmath6-braid word @xmath341 deleting all but the first @xmath342 in @xmath343 yields @xmath344 , which has @xmath345 as closure . if @xmath293 is @xmath346 for some positive integer @xmath331 , we write @xmath347 as the closure of @xmath348 deleting all but the first @xmath342 yields @xmath349 , which has closure @xmath350 . finally , if @xmath293 is @xmath351 for some positive integer @xmath331 , view @xmath352 as the closure of @xmath353 deleting all but the first @xmath342 yields @xmath354 , which has @xmath355 as closure . we view @xmath356 as the closure of the 4-braid @xmath357 . using the fact that the half twist on 4 strands @xmath358 anti - commutes with every other 4-braid , i.e. @xmath359 , @xmath360 , and @xmath361 , we have that @xmath362 with this we can write @xmath357 as follows . @xmath363 deleting the last @xmath331 occurrences of @xmath342 in this braid word gives @xmath364 which has closure @xmath365 similarly , the torus link @xmath366 is the closure of the @xmath1-braid @xmath367 deleting the last @xmath368 occurrences of @xmath342 yields a @xmath1-braid that has closure @xmath355 . we now study , which @xmath32 is subword - adjacent to @xmath369 and @xmath43 . our result is roughly that , whenever @xmath370 , then @xmath32 is subword - adjacent to @xmath369 and @xmath43 . our interest in this stems from the fact that this is an improvement over what is known in the algebraic setting ; compare with . in other words , the algebraic cobordism obtained by applying lemma [ lemma : algrealization ] to the subword - adjacencies provided in the following proposition is not known to come from an algebraic adjacency between @xmath371 and @xmath372 or a singular algebraic curve of degree @xmath293 with an @xmath373-singularity . [ rem : tmmtot2n ] after a first preprint of this article appeared , the author was pointed to work of orevkov , where the same bound is attained in a very similar setting , and his result was also motivated by questions discussed in section [ sec : motivation ] . indeed , orevkov s result ( * ? ? ? * theorem 3.13 ) allows to conclude that there exists an optimal cobordism between @xmath32 and @xmath369 , whenever @xmath370 . [ prop : subsurfacesoftmm ] let @xmath293 and @xmath93 be positive integers . if @xmath374 , then @xmath32 is subword - adjacent to @xmath369 . [ rem : subsurfacesoftm(m+1 ) ] a similar statement holds for the knots @xmath43 : let @xmath293 and @xmath93 be positive integers . if @xmath375 , then @xmath32 is subword - adjacent to @xmath43 . [ rem : subwordsoftmm ] we do not know whether the factor @xmath376 is optimal . if it is , the straight - forward application of @xmath3 does not suffice to show this . in fact , it only gives us that , whenever there is an optimal cobordism between @xmath43 and @xmath32 , then @xmath377 which is the same upper bound that is known for the algebraic setting ; see . indeed , let us fix a positive integer @xmath293 and assume that there exists an optimal cobordism between the @xmath32 and @xmath43 for some odd @xmath378 ; i.e. @xmath379 using @xmath380{oss_2014}},\ ] ] the obstruction given in yields @xmath381 a similar calculation using the signature instead of @xmath39 also yields . we proceed with the proof of proposition [ prop : subsurfacesoftmm ] , which can be adapted to yield remark [ rem : subsurfacesoftm(m+1 ) ] . we denote by @xmath382 the half twist on @xmath293 strands ; i.e. the @xmath293-braid @xmath383 the torus link @xmath369 is the closure of the full twist on @xmath293 strands @xmath384 . the main step in the proof consists of deleting generators in @xmath382 yielding a braid that is a split union of positive @xmath159-braids and which has roughly length @xmath385 . more precisely , we delete the generator @xmath386 in @xmath382 and then apply braid relations to get the positive braid word @xmath387 then , we delete all @xmath342 yielding a split union of @xmath388 on strands 1 and 2 , a half twist on the strands 3 to @xmath389 , and strand @xmath293 . we illustrate this for @xmath390.@xmath391!{\xcapv[6]@(0 ) } [ lu]!{\xcapv[6]@(0 ) } [ lu]!{\xcapv[6]@(0 ) } [ lu]!{\xcapv[6]@(0 ) } [ lu]!{\xcapv[6]@(0 ) } [ lu]!{\xcapv[6]@(0 ) } [ lu]!{\xcapv[6]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \color{red}\xcaph[1]@(0 ) } [ u(1.5)llllll ] ! { \color{black}\xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(1)lllll ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(0.5)llll ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ lll ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5)ll ] ! { \xcaph[1]@(0 ) } } \hspace{-3.8pc}\longrightarrow ~\xygraph { ! { 0;/r0.7pc/ : } [ u(3)]!{\xcapv[6]@(0 ) } [ lu]!{\xcapv[6]@(0 ) } [ lu]!{\xcapv[6]@(0 ) } [ lu]!{\xcapv[6]@(0 ) } [ lu]!{\xcapv[6]@(0 ) } [ lu]!{\xcapv[6]@(0 ) } [ lu]!{\xcapv[6]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(1)lllll ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(1)lllll ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(0.5)llll ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ lll ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5)ll ] ! { \xcaph[1]@(0 ) } } \hspace{-3.8pc}= ~\xygraph { ! { 0;/r0.7pc/ : } [ u(3.75)]!{\xcapv[7.5]@(0 ) } [ lu]!{\xcapv[7.5]@(0 ) } [ lu]!{\xcapv[7.5]@(0 ) } [ lu]!{\xcapv[7.5]@(0 ) } [ lu]!{\xcapv[7.5]@(0 ) } [ lu]!{\xcapv[7.5]@(0 ) } [ lu]!{\xcapv[7.5]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \color{red}\xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \color{black}\xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(1)lllll ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \color{red}\xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \color{black}\xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(0.5)llll ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \color{red}\xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \color{black}\xcaph[1]@(0 ) } [ lll ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \color{red}\xcaph[1]@(0 ) } [ d(0.5)ll ] ! { \color{black}\xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } } \hspace{-4.8pc}\longrightarrow ~\xygraph { ! { 0;/r0.7pc/ : } [ u(2.75)]!{\xcapv[5.5]@(0 ) } [ lu]!{\xcapv[5.5]@(0 ) } [ lu]!{\xcapv[5.5]@(0 ) } [ lu]!{\xcapv[5.5]@(0 ) } [ lu]!{\xcapv[5.5]@(0 ) } [ lu]!{\xcapv[5.5]@(0 ) } [ lu]!{\xcapv[5.5]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5)r ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(1)lllll ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5)r ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(0.5)llll ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5)r ] ! { \xcaph[1]@(0 ) } [ lll ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } } \hspace{-3.5pc},\ ] ] where arrows indicate the deletion of the generators marked in red . to the remaining half twist , which we readily identify with @xmath392 , we apply the same procedure . and we do this inductively until the remaining half twist is @xmath393 , or @xmath394 , where @xmath394 is just the trivial 1-strand braid . applying the procedure to @xmath395 just yields the split union of @xmath396 and one strand . on @xmath397 and @xmath394 it does not do anything . this inductive procedure yields a braid @xmath398 , which closes to a split union of torus links of braid index 2 . as before we illustrate this for @xmath390 . @xmath399!{\xcapv[6]@(0 ) } [ lu]!{\xcapv[6]@(0 ) } [ lu]!{\xcapv[6]@(0 ) } [ lu]!{\xcapv[6]@(0 ) } [ lu]!{\xcapv[6]@(0 ) } [ lu]!{\xcapv[6]@(0 ) } [ lu]!{\xcapv[6]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(1.5)llllll ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(1)lllll ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(0.5)llll ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ lll ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5)ll ] ! { \xcaph[1]@(0 ) } } \hspace{-3.8pc}\overset{\eqref{eq : delta71}}{\longrightarrow } ~\xygraph { ! { 0;/r0.7pc/ : } [ u(2.75)]!{\xcapv[5.5]@(0 ) } [ lu]!{\xcapv[5.5]@(0 ) } [ lu]!{\xcapv[5.5]@(0 ) } [ lu]!{\xcapv[5.5]@(0 ) } [ lu]!{\xcapv[5.5]@(0 ) } [ lu]!{\xcapv[5.5]@(0 ) } [ lu]!{\xcapv[5.5]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5)r ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \color{red}\xcaph[1]@(0 ) } [ u(1)lllll ] ! { \color{black}\xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5)r ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(0.5)llll ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5)r ] ! { \xcaph[1]@(0 ) } [ lll ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } } \hspace{-3.5pc}\longrightarrow\cdots\longrightarrow ~\xygraph { ! { 0;/r0.7pc/ : } [ u(2.75)]!{\xcapv[5.5]@(0 ) } [ lu]!{\xcapv[5.5]@(0 ) } [ lu]!{\xcapv[5.5]@(0 ) } [ lu]!{\xcapv[5.5]@(0 ) } [ lu]!{\xcapv[5.5]@(0 ) } [ lu]!{\xcapv[5.5]@(0 ) } [ lu]!{\xcapv[5.5]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ u(3)r ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } } \hspace{-3.5pc } = \beta_7\ ] ] the length @xmath400 of @xmath398 is described by the following formula . @xmath401 we use the above to obtain a braid @xmath402 that closes to a @xmath32 by deleting generators in @xmath384 , which shows that @xmath32 is subword - adjacent to @xmath369 . for this we write @xmath403 where @xmath404 is a half twist on the first @xmath405 strands . now , we apply the above deleting algorithm to @xmath404 , which is seen as @xmath406 , and @xmath382 yielding @xmath407 where @xmath408 is the @xmath293 strand braid which is obtained by having @xmath409 on the first @xmath405 strands . the braid @xmath402 is of the form @xmath410 where @xmath411 and @xmath412 are positive integers . as above we illustrate this for @xmath390 . @xmath413!{\xcapv[14]@(0 ) } [ lu]!{\xcapv[14]@(0 ) } [ lu]!{\xcapv[14]@(0 ) } [ lu]!{\xcapv[14]@(0 ) } [ lu]!{\xcapv[14]@(0 ) } [ lu]!{\xcapv[14]@(0 ) } [ lu]!{\xcapv[14]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(1.5)llllll ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(1)lllll ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(0.5)llll ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ lll ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5)ll ] ! { \xcaph[1]@(0 ) } [ ddl ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \color{red}\xcaph[1]@(0 ) } [ d(1.5 ) ] ! { \color{black}\xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(2.5)llllll ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(1.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(2)lllll ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(1.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(1.5)llll ] ! { \xcaph[1]@(0 ) } [ d(1.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ lll ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5)ll ] ! { \xcaph[1]@(0 ) } } \hspace{-8.1pc } \rightarrow ~\xygraph { ! { 0;/r0.6pc/ : } [ u(7)]!{\xcapv[14]@(0 ) } [ lu]!{\xcapv[14]@(0 ) } [ lu]!{\xcapv[14]@(0 ) } [ lu]!{\xcapv[14]@(0 ) } [ lu]!{\xcapv[14]@(0 ) } [ lu]!{\xcapv[14]@(0 ) } [ lu]!{\xcapv[14]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(1.5)llllll ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(1)lllll ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(0.5)llll ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ lll ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5)ll ] ! { \xcaph[1]@(0 ) } [ ddl ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(2)r ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(2.5)llllll ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(1.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(2)lllll ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(1.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(1.5)llll ] ! { \xcaph[1]@(0 ) } [ d(1.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ lll ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5)ll ] ! { \xcaph[1]@(0 ) } } \hspace{-8.1pc}= ~\xygraph { ! { 0;/r0.6pc/ : } [ u(7.25)]!{\xcapv[14.5]@(0 ) } [ lu]!{\xcapv[14.5]@(0 ) } [ lu]!{\xcapv[14.5]@(0 ) } [ lu]!{\xcapv[14.5]@(0 ) } [ lu]!{\xcapv[14.5]@(0 ) } [ lu]!{\xcapv[14.5]@(0 ) } [ lu]!{\xcapv[14.5]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(1.5)llllll ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(1)lllll ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(0.5)llll ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ lll ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5)ll ] ! { \xcaph[1]@(0 ) } [ d(1.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \color{red}\xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \color{black}\xcaph[1]@(0 ) } [ d(2)r ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(2.5)llllll ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \color{red}\xcaph[1]@(0 ) } [ d(1.5)r ] ! { \color{black}\xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(1.5)lllll ] ! { \xcaph[1]@(0 ) } [ d(2)r ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(2)llll ] ! { \xcaph[1]@(0 ) } [ d(2 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ lll ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5)ll ] ! { \xcaph[1]@(0 ) } } \hspace{-8.4pc } % \\ & \rightarrow ~\xygraph { ! { 0;/r0.6pc/ : } [ u(6.75)]!{\xcapv[13.5]@(0 ) } [ lu]!{\xcapv[13.5]@(0 ) } [ lu]!{\xcapv[13.5]@(0 ) } [ lu]!{\xcapv[13.5]@(0 ) } [ lu]!{\xcapv[13.5]@(0 ) } [ lu]!{\xcapv[13.5]@(0 ) } [ lu]!{\xcapv[13.5]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(1.5)llllll ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(1)lllll ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(0.5)llll ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ lll ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5)ll ] ! { \xcaph[1]@(0 ) } [ d(1.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5)r ] ! { \xcaph[1]@(0 ) } [ d(2)r ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(2.5)llllll ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(2)rr ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(2)lllll ] ! { \xcaph[1]@(0 ) } [ d(2)r ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(2)llll ] ! { \xcaph[1]@(0 ) } [ d(2 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ lll ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5)ll ] ! { \xcaph[1]@(0 ) } } \hspace{-7.9pc } \overset{\eqref{eq : delta72}}{\rightarrow } ~\hspace{-0.2pc}\xygraph { ! { 0;/r0.6pc/ : } [ u(6.75)]!{\xcapv[13.5]@(0 ) } [ lu]!{\xcapv[13.5]@(0 ) } [ lu]!{\xcapv[13.5]@(0 ) } [ lu]!{\xcapv[13.5]@(0 ) } [ lu]!{\xcapv[13.5]@(0 ) } [ lu]!{\xcapv[13.5]@(0 ) } [ lu]!{\xcapv[13.5]@(0 ) } [ u(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ u(3)r ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(3.5)lll ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(0.5)r ] ! { \xcaph[1]@(0 ) } [ d(2)r ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(2.5)llllll ] ! { \xcaph[1]@(0 ) } [ d(0.5)l ] ! { \xcaph[1]@(0 ) } [ d(2)rr ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(2)lllll ] ! { \xcaph[1]@(0 ) } [ d(2)r ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ u(2)llll ] ! { \xcaph[1]@(0 ) } [ d(2 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ lll ] ! { \xcaph[1]@(0 ) } [ d(0.5 ) ] ! { \xcaph[1]@(0 ) } [ d(0.5)ll ] ! { \xcaph[1]@(0 ) } } \hspace{-8pc}=\gamma_7\end{aligned}\ ] ] the closure of @xmath402 is a braid index two torus link @xmath32 . this follows from observing that the closure of @xmath414 is @xmath415 . since @xmath416 we see that @xmath417 ; i.e. the closure of @xmath402 is @xmath418 . with the above calculations for @xmath400 we get @xmath419 for @xmath330 , @xmath420 , and @xmath421 , respectively this finishes the proof since @xmath93 is the largest integer with @xmath374 . for completeness , we provide the calculations that yield the @xmath39-values given in . more generally , the calculation we provide below in the proof of proposition [ prop : upsilonfort3t4 ] allows to determine the function @xmath423\to\r$ ] for torus knots @xmath19 of braid index 3 and 4 : for all positive integers @xmath93 , we have @xmath424 this fully describes @xmath425 since it is symmetric , i.e. @xmath426 ; see @xcite . our calculations have convinced us that , for a general torus knot @xmath64 , @xmath427 might look similar to the homogenization of the signature profile of torus knots ; i.e. the following function , studied by gambaudo and ghys @xcite : @xmath428\to\r,\quad t\mapsto \lim_{k\to\infty}\frac{\s_{e^{\pi it}}\left(\overline{((a_1\cdots a_{p-1})^q)^k}\right)}{k},\ ] ] where @xmath429 denotes the levine - tristram signature . we hope to explore this further in the future . the only heegaard - floer theory input in the proof of proposition [ prop : upsilonfort3t4 ] is the following combinatorial procedure to determine @xmath3 for torus knots ( or more generally @xmath77-space knots ) @xcite : write the alexander polynomial @xmath430 as @xmath431 , where @xmath432 is a decreasing sequence of integers . construct a corresponding decreasing sequence of integers @xmath433 defined by @xmath434 then one has @xmath435{oss_2014}}.\ ] ] in particular , @xmath436 in fact , for the calculation one only needs the evenly indexed @xmath437 , for which can be shortened to @xmath438 for @xmath454 , we calculate first when @xmath93 is even , i.e. @xmath455 for a positive integer @xmath81 : @xmath456 which means @xmath457 therefore , we have @xmath458which yields that @xmath459 equals @xmath460 finally , for @xmath93 odd , a similar calculation yields @xmath461 which means @xmath462 , @xmath463@xmath464 therefore , we have @xmath465 this yields that @xmath459 equals @xmath466 lars v. ahlfors . . mcgraw - hill book co. , new york , third edition , 1978 . an introduction to the theory of analytic functions of one complex variable , international series in pure and applied mathematics . s. yu . orevkov . some examples of real algebraic and real pseudoholomorphic curves . in _ perspectives in analysis , geometry , and topology _ , volume 296 of _ progr . _ , pages 355387 . birkhuser / springer , new york , 2012 . john r. stallings . constructions of fibred knots and links . in _ algebraic and geometric topology ( proc . pure math . , stanford univ . , stanford , calif . , 1976 ) , part 2 _ , proc . pure math . , xxxii , pages 5560 . amer . math . soc . , providence , r.i .	we construct cobordisms of small genus between torus knots and use them to determine the cobordism distance between torus knots of small braid index . in fact , the cobordisms we construct arise as the intersection of a smooth algebraic curve in @xmath0 with the unit 4-ball from which a @xmath1-ball of smaller radius is removed . connections to the realization problem of @xmath2-singularities on algebraic plane curves and the adjacency problem for plane curve singularities are discussed . to obstruct the existence of cobordisms , we use ozsvth , stipsicz , and szab s @xmath3-invariant , which we provide explicitly for torus knots of braid index 3 and 4 .
local damage of buildings can either be due to accidental events like gas explosions , gross design - construction errors and malicious terrorist attacks @xcite , or can be thoroughly planned as part of controlled demolition processes with blast . subsequent cascades of failures can cause large economic and human loss when triggered by accidental events and make the difference between effective and dangerously ineffective demolitions . research about progressive collapse of buildings proceeded discontinuously since 1970s mostly prompted by outstanding and shocking catastrophes . interest in the subject rose after the ronan point partial collapse in 1968 due to gas explosion . during the seventies , the fundamental approaches to structural robustness as well as many indicators , like the reserve strength ratio ( rsr ) @xcite , were formulated , also with regard to off - shore structures that suffered brittle collapses on the north sea . renewed attention to the problem was given due to terrorist attacks against the alfred p. murrah federal building , oklahoma city 1995 , and the world trade center ( wtc ) , new york 2001 @xcite . nowadays many codes prescribe alternate paths for the load ( alternate load path method ( alpm ) @xcite ) and high toughness of structural members and their interconnections @xcite . nonetheless , these measures are not always sufficient to prevent progressive collapse @xcite . moreover , even though the serious damage amplification due to dynamics has been exhaustively pointed out in @xcite , static analyses are still used in the context of the alpm . developing efficient tools to evaluate structural robustness and to prove the effectiveness of measures aimed at preventing progressive collapse is therefore an important issue and today several algorithms and models are available in literature . a simplified approach to take dynamics and impacts between falling elements into account was proposed by @xcite and @xcite showed its analogy with the variational approach to fracture mechanics . the scheme is based on energy balance , requires only static analyses , and was effectively applied to strain hardening @xcite and softening @xcite structural elements . analytical 1d models were developed after the wtc collapse @xcite . in these models , progressive buckling of the columns is due to the impact of the upper floors , considered as an increasing falling mass . differently , computer simulations permit to study more complex 2d and 3d structures . several key factors that influence robustness of frames have already been identified . for instance , we know that the loss of external columns from the facades or the corners of a buildings are the most serious scenarios where , according to the alpm method , one column is instantaneously removed ( see , e.g. @xcite ) . moreover , it was shown that beam - column connections are critical points of failure initiation @xcite and that catenary effects in the floor slabs can remarkably improve robustness @xcite . even though the final outcome of progressive collapse depends on the collisions between structural elements , most of literature focuses on collapse initiation . collisions are rarely taken into account either detailed with finite elements @xcite , or approximated in the framework of finite macro - elements @xcite . detailed finite elements are too demanding in terms of computational time for extensive parametric studies on large structures . differently , finite macro - elements are efficient and can be applied to large structures , but they require strong approximations to take into account collisions and catenary effects , especially in 3d ( e.g. see @xcite ) . the lack of experimental results of progressive collapses suggests an approached based on simulations whose reliability arises from the basic physics incorporated . the results obtained with such algorithms can be used to construct , test and calibrate simpler models . in this work , we use spherical discrete elements ( de ) to simulate the progressive collapse of typical 3d framed structures made of reinforced concrete with fixed regular overall geometry ( see sec . [ simul ] ) . the aim is to study the collapse initiation mechanisms due to dynamic stress redistribution , and the subsequent damage propagation mechanisms due to collisions between the structural elements . understanding the activated mechanisms , depending on the strength , the stiffness and the plastic properties of the structural elements , can help to choose optimal robustness oriented design solutions as well as the most appropriate structural reinforcement of existing buildings . we perform parametric studies scaling the cross sectional size and reinforcement by the _ cross sectional scale factor _ @xmath0 and varying the plastic capacity @xmath1 of the structural elements ( see sec . [ sims ] ) . in this way , we show the expected collapse mechanisms and the final consequences of progressive collapse in terms of final collapse extent and fragment size distribution for various ( @xmath0 , @xmath1 ) . finally , in sec . [ rcgen ] we compare the damage response of structures with symmetric and asymmetric reinforcement in the beams . the choice of spherical de as simulation tool is motivated by several factors @xcite : first of all , de are naturally suitable to deal with dynamic problems since they are based on the direct integration of newton s equations of motion , which makes the algorithm simple and fast . moreover , geometric and material nonlinearities , as well as local ruptures can be easily modelled without remeshing . momentum transmissions due to collisions can be included straightforwardly ( see sec . [ mdmodel ] ) . a quite fine mesh is required to represent the actual volume of the structure and to reduce the error originating from the fact that instead of considering sectional ruptures we instantly remove beam elements that are responsible for the cohesion of the system ( see sec . [ mdmodel ] ) . considering sectional ruptures would require remeshing while a more precise representation of the volumes could be obtained with polyhedrical de . for both strategies , the computational demand would grow remarkably . our model has previously been employed to study fragmentation of materials , e.g. @xcite , and its applicability to progressive collapse of structures is demonstrated in @xcite . in the simulations , the intact structures are first equilibrated under external service load and gravity . if some elements fail during this initial phase , the structure is incapable of carrying the service load and the simulation is stopped . differently , if no elements fail , the local damage induced by an accidental event is considered by a sudden removal of a central column of a facade at the first floor ( fig . [ figmdstruct ] ) , according to the alpm . the subsequent dynamic stress redistribution can break other elements and trigger widespread progressive collapse . the dynamics of the system is followed by means of explicit time integration , using a 5@xmath2 order gear predictor - corrector scheme that , for the explored set of parameters , is stable with time increments lower than @xmath3 seconds . in the following subsections [ struct ] and [ mdmodel ] we give a detailed description of the studied structures and employed model . . c ) cross sections of the structural elements and arrangement of the reinforcement . ] for comparative reasons we limit ourselves to study typical regular 3d frames formed by 4x4x4 identical square cuboid cells with @xmath4=4 m and @xmath5=3 m ( see fig . [ figstruct].a ) . the structures are made of columns along the vertical @xmath6 direction , clamped to the ground at @xmath7 m and connected at each storey by principal beams in @xmath8 and @xmath9 direction . thin slabs spanning between the principal beams form the floors while the presence of walls is not considered . the geometry of the cross sections of the structural elements is displayed in fig . [ figstruct].c , where the subscripts @xmath10 , @xmath11 and @xmath12 denote columns , beams , and slabs . we set the height of the cross sections proportional to the length of the structural element with the coefficients @xmath13 , namely @xmath14=h/10 , @xmath15=l/10 and @xmath16=l/50 , and we scale each @xmath17 by a dimensionless _ cross sectional scale factor _ @xmath0 . @xmath0 is identical for all elements and enlarges their cross sections making the structure stiffer and stronger . the base edges @xmath18 , @xmath19 of the cross section are proportional to the heights @xmath15 , @xmath19 with the aspect ratio coefficients @xmath20=2/3 and @xmath21=1 . consequently , the area , the sectional inertia with respect to the @xmath22 principal direction and the torsional inertia of the cross sections of beams , columns and portions of floor slabs can be easily computed . we represent a structure made of reinforced concrete ( rc ) with young s modulus @xmath23 and shear modulus @xmath24 ( see appendix , table [ tabmecpar ] ) . the reinforcement is symmetrically distributed ( see fig . [ figstruct].c ) and its area is proportional to that of the cross section by the factors @xmath25=1.78% , @xmath26=0.58% , and @xmath27=1.26% . the structure carries its own weight @xmath28 , the service external dead load @xmath29=285kg / m@xmath30 given by non structural elements like pavement , plaster and internal walls , and the service live load @xmath31=667n / m@xmath30 . @xmath29 and @xmath31 are considered uniformly applied to the upper faces of the floors . we represent the columns and the beams by meshes with @xmath32 and @xmath33 euler - bernoulli ( eb ) beam elements respectively . the floor slabs are considered by a grid of @xmath34 eb elements of length @xmath35 ( see fig . [ figmdstruct].b ) that define slab portions of size @xmath36 ( see fig . [ figstruct].c ) . to represent the volume of the structure , we set @xmath33 and @xmath32 to obtain @xmath37 , where @xmath38 and @xmath39 are the height of the cross section along @xmath40 and the length of the generic eb element . the cross sections of the eb beam elements are set according to sec . [ struct ] . the error on @xmath41 introduced by the simplifying hypothesis @xmath42 is acceptable because the bending of the beams and of the slabs in the horizontal plane is not relevant . the light grey area marks the initially removed column . generic euler - bernoulli element in c ) undeformed and d ) deformed state . ] we represent the structural volume by spheres surrounding each node ( see fig . [ figmdstruct].a ) . the diameter of the @xmath43 sphere is equal to 90@xmath44 of the length of the shortest eb element connected to node @xmath45 . the mass @xmath46 is obtained summing the contributions from the mass @xmath47 of the @xmath48 eb elements connected to node @xmath45 and from an extra mass @xmath49 given by the external dead load @xmath29 : @xmath50 note that in the de algorithm the eb beam elements do not have a mass since @xmath47 is concentrated in the spheres . the rotational inertia of a sphere is computed considering @xmath46 to be uniformly distributed . the eb elements determine the interactions between pairs of nodes , associating their relative rotations and displacements to forces and moments acting on them . in the following , we describe the linear elastic - perfectly plastic constitutive behavior and the failure rules of the eb elements . [ [ elastic - regime ] ] elastic regime + + + + + + + + + + + + + + in the linear elastic regime , we use the force - displacement relations described in detail in @xcite , thus taking into account the geometric nonlinearities due to large displacements and neglecting shear deformability . the rotations @xmath51 are defined starting from the deformed state of the generic eb element ( see fig . [ figmdstruct].d ) . namely , the bending rotations @xmath52 , @xmath53 around the @xmath22 , @xmath40 principal axes align the @xmath54 , @xmath55 axes with the line connecting nodes 0 and 1 , while rotating the @xmath56 , @xmath57 axes around @xmath58 by @xmath59 makes them parallel to @xmath60 , @xmath61 : @xmath62 at a given time step , we compute the axial strain @xmath63 and the rotations @xmath51 for every eb element starting from the position of the spheres and at the orientation of the @xmath22 , @xmath40 , @xmath58 axes frozen to nodes 0 and 1 . the forces and moments at nodes 0 and 1 come from the eb beam theory : @xmath64 @xmath23 is the young s modulus of concrete ( see table [ tabmecpar ] ) . @xmath65 and @xmath66 denote the bending moment and the effective rotation around the @xmath67 axis at node @xmath68 , @xmath69 is the shear force along the @xmath70 axis at node @xmath71 , @xmath72 is the normal force and @xmath73 is the torsion . damping inside the beams is considered adding forces and moments at nodes 0 and 1 that are proportional to the elastic part of the velocities @xmath74 , @xmath75 , @xmath76 , @xmath77 , @xmath78 , @xmath79 by the coefficients in table [ taby - th - damp ] but with opposing direction @xcite . [ [ plastic - regime ] ] plastic regime + + + + + + + + + + + + + + progressive collapse of structures involves many irreversible processes in the elements and plastic energy dissipation can determine robust or vulnerable responses to damage . to consider plasticity in the eb elements , we make the simplifying assumption that axial and bending plasticization are uncoupled . this choice is justified by our intention to keep the model as simple as possible , leaving refinements to further works . furthermore , we do nt consider plasticization due to shear or torsion in the rc because plastic dissipations associated to them are generally small . under these hypotheses , the eb elements enter the perfectly plastic regime in axial direction or in bending at node @xmath71 and around @xmath22 or @xmath40 independently if one or more of these conditions is satisfied : @xmath80 we set the tension and compression yield thresholds @xmath81 , @xmath82 neglecting the contribution of concrete in tension and steel in compression . the bending yield threshold @xmath83 is evaluated referring to the @xmath22 axis and neglecting the contribution of concrete . if @xmath84 is the fraction of reinforcement in tension , i.e. @xmath85 for the columns and @xmath86 for the beams and the slabs ( see fig . [ figstruct].c ) , we obtain : @xmath87 we also add a further contribution @xmath88 to @xmath89 to consider that compressions @xmath90 increase @xmath89 by reducing the area of concrete in tension during bending . we set @xmath88 assuming that it is carried by the reinforcement alone and that it compensates the strain @xmath91 in the reinforcement in tension produced by @xmath72 , i.e. @xmath92 : @xmath93 within the employed direct time integration scheme , we check yielding in terms of strain and rotations instead of forces and moments . thus , we adopt elongation @xmath94 , shortening @xmath95 and bending @xmath96 yield thresholds that satisfy eqs . [ ny_cond ] - [ by - cond ] in the equality form , when inserted into eqs . [ m - fi ] - [ mz - teta ] . the expressions for the @xmath94 , @xmath95 and @xmath96 are summarized in table [ taby - th - damp ] , where also the additional term to @xmath96 from eq . [ dmy ] is shown . in the perfectly plastic regime , we consider the axial strain @xmath97 and rotations @xmath98 , @xmath99 to result from the sum of an elastic and a plastic contribution . [ figfluxey ] shows how axial plasticization is implemented at the generic time step @xmath100 . history dependence is considered accumulating the plastic strain @xmath101 in time . the linear distribution of bending along the eb element makes the description of the rotational plastic regime more complicated . first , we compare @xmath102 , @xmath103 obtained from the integration of newton s equation with @xmath96 to define whether plasticization occurs only at one node or at both . if for instance only node 0 enters the plastic regime , then we set @xmath104 = @xmath105 + @xmath106 , where @xmath105 , put into eq . [ m - fi ] , satisfies eq . [ by - cond ] in the equality form . differently , if both node 0 and 1 plasticize , either @xmath104 and @xmath107 must be separated into elastic and plastic part , so that eq . [ m - fi ] and eq . [ by - cond ] in the equality form return a linear system of two equations in the two unknowns @xmath105 and @xmath108 . as for axial plasticization , the plastic parts @xmath106 and @xmath109 must be subtracted from the total @xmath104 and @xmath107 at the next time step , and the plastic rotations must be cumulated in time . [ [ element - failure ] ] element failure + + + + + + + + + + + + + + + if the strain in a cross section of an eb elements is too high , the element fails and is instantly removed from the system . in the following , we will scale the plastic capacity of the eb elements by a parameter @xmath1100 . if @xmath111 the elements break when a combination of the elongation and the effective rotations at nodes @xmath68 is large with respect to the yielding thresholds evaluated in uncoupled conditions : @xmath112 differently , if @xmath113 , the breaking rules are : @xmath114 with ultimate threshold values of elongation @xmath115 , shortening @xmath116 and rotation @xmath117 estimated in uncoupled conditions . failure due to shear and torsion is neglected because the shear reinforcement is thought to be designed according to the capacity design approach , that avoids the occurrence of these brittle mechanisms before bending or axial strain failure . we assume @xmath115 and @xmath116 equal to the ultimate tensile strain of steel and compressive strain of concrete respectively ( see table [ tabmecpar ] ) . @xmath117 is estimated considering a state of uniform bending , and thus uniform curvature @xmath118 , in the eb element . under these hypotheses , the rotation between the edges of an eb element is @xmath119 and gives a strain in the steel bars at @xmath40 = @xmath120 equal to @xmath121 . we thus obtain @xmath117 setting @xmath91 equal to the ultimate strain of the steel , i.e. @xmath122 . the expressions in table [ taby - th - damp ] consider also that the adopted meshing rule assures @xmath123 . progressive failure of eb elements can lead to the free fall of structural elements . we model inter - spheres collisions by a hertzian potential @xcite that generates a conservative repulsive force between partially overlapping spheres . this force is directed along the line connecting the centres of mass of the colliding spheres and is proportional to the overlapping volume by a stiffness parameter @xmath9 . a similar force is generated when a sphere crosses the @xmath7 plane that represents the ground . impacts dissipates energy due to local fragmentation and to sliding and rolling friction . thus we introduce forces that are proportional and opposed to the normal , the tangential and the rotational relative velocities of the overlapping spheres with the damping coefficients summarized in table [ tabmdimpacts ] . in tangential direction , either coulomb s or dynamic sliding friction are considered . the plastic capacity of the structural elements is a key factor of progressive collapse . plasticity determines the subsequent mechanisms of damage propagation based on collisions and the final extent of collapse . we employ the presented model to perform parametric studies on the structural _ cross sectional scale factor _ @xmath0 ( see sec . [ struct ] ) and on the _ plastic capacity _ @xmath1 ( see eqs . [ tbreak ] and [ cbreak ] ) . we choose @xmath0 as parameter because a structure with given @xmath1 can be robust or extremely vulnerable to a column removal , depending on the cross sectional size of the elements . in this way , we can see the effect of plasticity in structures that exhibit different responses to the initial damage , ranging from no collapse to catastrophic collapse . in sec . [ colmec ] we describe the observed global and local primary mechanisms that can trigger progressive collapse and the subsequent collisions - driven mechanisms . in sec . [ collapse ] we show the results of the parametric studies . we especially point out what collapse mechanisms occur depending on @xmath0 and @xmath1 and what are the consequences in terms of final extent of the collapse . finally , in sec . [ frag ] we show how @xmath0 and @xmath1 effect the fragment size distribution of the rubble . the initial column removal can trigger three different primary collapse mechanisms that start progressive collapse , two of which are global and lead to total collapse : 1 . the first one is caused by elastic waves inside the floors that can separate even distant slabs from the beams ( see fig . [ figglobcoll].a ) . + , @xmath124 ) and b ) progressive punching ( @xmath125 , @xmath126 ) . local starting collapse ( @xmath127 , @xmath124 ) : a ) detail of the first failure area , where the light gray area marks the initially removed column and the arrows show the direction of crack propagation ; b ) first stages of the local progressive collapse . e ) approximate static schemes of the elements where the starting damage propagation occurs ; the springs in b represent the stiffness of the perpendicular beams . ] 2 . the other global mechanism separates less slabs from the beams but the floors are progressively punched by the columns ( see fig . [ figglobcoll].b ) . most probably , this mechanism would turn into progressive buckling of the columns if they were less reinforced . the local primary collapse mechanism is characterized by a crack propagating from point a ( fig . [ figglobcoll].c ) and disconnecting the neighboring floor slabs from the beams ( fig . [ figglobcoll].d ) . + to explain why rupture occurs at point a instead of point b or c we consider the schematic representation in fig . [ figglobcoll].e . after the column removal , the cross sections @xmath128 and @xmath129 , being at two sides of the same node b , show the same vertical displacement . in fig . [ figglobcoll].e , the load on @xmath130 on beam @xmath131 is greater than the load @xmath132 on beam @xmath133 because the area that can transfer load to the former is larger . therefore , considering that the torsional stiffness of @xmath133 can be represented by a torsional spring that reduces the bending moment in @xmath128 and increases that in @xmath134 , the maximum static bending moment is located in @xmath134 . when a local primary collapse mechanism is triggered , the portion of structure above the removed column undergoes free fall and collides with the floor slabs below . this @xmath135 effect ( fig . [ fighamm ] ) generates elastic waves that can damage the neighbouring floor slabs . rarely does it have catastrophic consequences by itself . , @xmath124 ) . free fall of the floor slabs , c ) almost horizontal for brittle structures ( @xmath0 = 0.59 , @xmath1 = 0.2 ) and d ) tilted for plastic structures ( @xmath136 , @xmath126 ) . in brittle structures e ) the slabs stacking on the ground push less against the column than in f ) plastic structures ] other sources of damage transmission are the lateral impacts between falling rubble and still intact portions of the structure . this @xmath137 effect can destroy the perimeter beams of neighbouring floor slabs , which can eventually collapse , but usually is not able to cause a widespread propagation of damage by itself . finally , a severe secondary mechanism is due to the forces exerted by the rubble stacking on the ground , which can cut the neighbouring columns at the base ( @xmath138 , see fig . [ fighamm].c - f ) . the collapse mechanisms described in sec . [ colmec ] can occur depending on the cross sectional size and reinforcement of the structural elements , determined by the cross sectional scale factor @xmath0 , and on the available plastic capacity @xmath1 . fig . [ figphases].a shows the response of the studied frames to the applied local damage . [ htb ] is the maximum plastic rotation in uncoupled condition ( cf . [ tbreak ] , [ cbreak ] ) ] for ( @xmath0 , @xmath1 ) pairs below the @xmath139 curve , the intact structures experience static collapse before the damage . such weak frames are not supposed to exist since they can not carry the service load . structures with ( @xmath0 , @xmath1 ) within the dashed area just above the @xmath139 curve completely collapse after the column removal triggers a global primary mechanism . this ( @xmath0 , @xmath1 ) region is narrow and vanishes when @xmath140 since brittle failures induce compartmentalisation , abruptly interrupting the dynamic stress flow . in this way , either the propagation of waves ( dominating when @xmath141 ) and the global failure of the storeys due to progressive punching ( @xmath142 ) are avoided . structures with ( @xmath0 , @xmath1 ) above the dashed region and below the @xmath143 exhibit local primary collapse , but can still experience total collapse prompted by the @xmath138 mechanism ( see fig . [ figphases].b ) . if @xmath1 is small , the probability that a sequence of @xmath138 provokes total collapse is low , due to the almost null tilt of the floor slabs during the free fall , the high degree of fragmentation after collisions and the larger @xmath0 values . on the contrary , when @xmath1 is large , sequences of @xmath138 are frequent because the slabs tilt while falling and stack on the ground without considerably fragmenting . consequently , massive slabs lean against the thin ( small @xmath0 ) base columns at a remarkable height , and cut them ( fig . [ fighamm].c - e ) . partial collapse can occur when @xmath144 and @xmath0 is sufficiently large ( see fig . [ figphases].b ) . in this case all the collision mechanisms of sec . [ colmec ] determine the final extent of the collapse . overlapping between the partial and total collapse regions in fig . [ figphases ] is due to the fact that the consequences of collision driven mechanism are considerably variable . structures with ( @xmath145 ) pairs above the @xmath143 curve are perfectly robust , i.e. they do not suffer any further failure after the column removal . finally , when @xmath146 the intact structure does not plasticize in the pre - dmage stage , i.e. during the static application of the service load . since this is a common requirement for buildings in service conditions , structures with realistic size of the elements are located in the @xmath147 region . , @xmath1 ) pairs between @xmath139 and @xmath143 . note that for each @xmath1 , values of @xmath0 larger or smaller than those of the plotted points give respectively @xmath148 and @xmath149 . b fragment mass distribution for different plastic capacities @xmath1 . @xmath150 is the fraction of fragments with mass @xmath48 normalised to the total mass of the structure @xmath151 . ] the beneficial effect of plasticity is evident from the fact that @xmath143 and @xmath139 decrease with @xmath1 ( see fig . [ figphases].a ) . these curves sensibly decrease in the range @xmath152 , which means that the provision of a minimal plastic capacity is sufficient to remarkably improve the structural response , assuring complete safety to structures with @xmath153 . this result is a consequence of the symmetric distribution of reinforcement inside the beams and the floor slabs , that makes their bending behavior qualitatively similar to that of steel elements . steel structures are actually likely to sustain one column removal without damage propagation . in sec . [ rcgen ] we show that rc structures with realistic cross sectional size of the elements and asymmetric reinforcement distribution would experience partial collapse . differently , when @xmath154 , @xmath143 and @xmath143 do not decrease much anymore ( see fig . [ figphases].a ) . this means that the initiation of progressive collapse is a local phenomenon associated with relatively small plastic stress redistributions , as already observed for 2d steel frames by @xcite . [ figcol - alp].a shows a quantitative measure of the final collapse extent for structures with different plastic capacity @xmath1 . in particular , we compute the _ demolition ratio _ @xmath155 , i.e. the fraction of lost living space at the end of the collapse . from the figure it can be immediately seen that structures with high plastic capacity @xmath1 undergo progressive collapse only if they have thin elements , i.e. small @xmath0 , but if the collapse is triggered it will affect the whole system . the large variability of @xmath155 is due to the collision - driven mechanisms and , in particular , the largest values of @xmath155 associated with partial collapse , i.e. when @xmath156 and @xmath157 , indicate the occurrence of @xmath158 . the size of the fragments produced after a structural collapse is interesting for controlled demolitions , where large fragments require further effort to relocate them . the probability density distribution @xmath159 of the mass of the fragments @xmath48 normalised by the total mass of the structure @xmath151 is shown in fig . [ figcol - alp].b . we observed that @xmath0 does not influence @xmath150 . differently , as the plastic capacity of the elements @xmath1 grows the power law regression lines in figure shift towards larger sizes of the fragments while their exponent , between -1.2 and -1.4 , does not seem to change according to a trend . note that this exponent is in the range of the one reported for shell fragmentation of 1.35 @xcite . when @xmath1 is small @xmath150 has large dispersion and most of the fragments are represented by single spheres completely disconnected from the others . calling @xmath160 the sum of the masses of the fragments made of one sphere and @xmath161 the total mass of the fragments , @xmath162 is larger than 0.7 when @xmath111 . this denotes a finite size effect , i.e. the single sphere is larger than the characteristic size of the fragments . on the contrary , when @xmath1 is high @xmath150 is less dispersed and @xmath163 implies that the fragment size distribution is better caught . .c , d ) symmetric and asymmetric reinforcement arrangement . static deformed condition of the beams , bending moment in @xmath131 and regions in tension b ) before and c ) after the column removal . ] the results in sec . [ collapse ] refer to rc structures with symmetrically reinforced elements . nevertheless , the reinforcement inside the beams and the floor slabs of real structures is mostly concentrated in the regions under tension in service conditions . fig . [ figasym].a shows the symmetric and the more realistic asymmetric reinforcement arrangement inside the beams that are involved in the local primary collapse mechanism of fig . [ figglobcoll].c , d . the column loss in fig . [ figasym].c produces the inversion of the bending moment in @xmath128 and thus , in case of asymmetric reinforcement , the reinforcement under tension @xmath164 in @xmath128 passes from @xmath165 before the damage ( see fig . [ figasym].b ) to @xmath166 after the damage ( see fig . [ figasym].c ) . therefore , if @xmath167 is small , section @xmath128 is likely to fail before section @xmath134 , differently from the observations in case of symmetric reinforcement ( see sec . [ colmec ] ) . in the following , depending on @xmath167 , we estimate the @xmath143 curves that separates the collapse and robustness regions in fig . [ figphases].a for frames with asymmetric reinforcement . beam ( cf . [ q(alfa ) ] ) . b ) transition to no collapse after the column removal for different ratios of asymmetry @xmath168 of the reinforcement in the beams . ] failure in @xmath128 occurs if : @xmath169 where @xmath170 is the static bending moment in @xmath128 after the damage ( see fig . [ figasym].c ) and @xmath171 is an amplifying factor that considers dynamics and plastic capacity . @xmath172 ranges from @xmath86 for linear elastic - perfectly brittle structures ( @xmath111 ) to @xmath173 for perfectly plastic structures with @xmath174 . @xmath175 is the load per unit length on the beam ( see fig . [ figload].a ) : @xmath176\varpi(\beta ) l \;\;,\ ] ] @xmath177 is the yield bending moment in @xmath128 , with @xmath178 because it is proportional to the area of the cross section times its height ( see eq . thus , eq . [ bba = byba ] can be rewritten as . @xmath179\varpi(\beta ) \frac{l^3}{6 } = dyn\cdot\psi t_s\rho_{s , b}\lambda_b^3\delta_b f_y \alpha^3 \;\;.\ ] ] for a given @xmath167 , solving eq . [ alfac ] in @xmath0 for different @xmath1 permits to trace the @xmath180 curves . nevertheless , in eq . [ alfac ] the parameter @xmath181 is still unknown . in order to assign values to @xmath181 , we repeat the previous argument for structures with symmetric reinforcement , for which we now the @xmath180 values at some @xmath1 obtained from our simulations ( see fig . [ figphases].a ) . in case of symmetric reinforcement , the @xmath131 beam in fig . [ figasym].c fails in @xmath134 , where the bending moment in @xmath182 . since in this case @xmath183 , eq . [ alfac ] turns into : @xmath184\varpi(\beta ) \frac{l^3}{3 } = dyn\cdot t_s\rho_{s , b}\lambda_b^3\delta_b f_y \alpha^3 \;\;.\ ] ] solving eq . [ alfac_sym ] starting from the @xmath180 values in fig . [ figphases].a , we obtain the @xmath181 values in table [ dyn_omega ] . .@xmath185 factors of eqs . [ alfac ] and [ alfac_sym ] . [ cols="<,^,^,^,^,^",options="header " , ] inserting the @xmath181 values from table [ dyn_omega ] into eq . [ alfac ] , we obtain the curves in fig . [ figload].b . these curves show that collapse can initiate also in well designed rc structures , i.e. with @xmath186 . moreover , we argue that the final extent of the collapse should be partial . note that the @xmath187 curve in fig . [ figphases ] does not change with @xmath168 since it only depends on the reinforcement in tension in the intact structure . also , the bold dotted line in fig . [ figphases].b regarding the base - cutting phenomenon does not depend on @xmath168 but only on the reinforcement inside the columns , that is usually symmetric . therefore , when @xmath188 an expansion of the partial collapse region towards higher values of @xmath0 is expected . progressive collapse of framed structures after local damage consist in an initial triggering and a subsequent damage propagation . if the initial damage is small , like the studied column removal , collapse initiation is generally a local phenomenon affecting the surroundings of the initially damaged area . global primary mechanisms can occur only in thin structures with enough plastic capacity to avoid the compartmentalisation effect produced by brittle ruptures , i.e. @xmath189 . nevertheless , if the starting damage is more serious than a single column removal , global primary mechanisms can be expected also for larger and more brittle structures . in case of multiple column removal , progressive crushing or buckling of the columns is a possible global primary collapse mechanism that was not observed in this context . we showed that structures with minimal plastic capacity and symmetrically reinforced beams are robust towards a single column removal . on the contrary , frames with asymmetric reinforcement would experience partial collapse even if made of elements with large plastic capacity . if a local primary collapse mechanism is triggered , the final extent of the collapse depends on secondary mechanisms driven by collisions between the structural elements . we showed that damage can not widely propagate in structures with small plastic capacity since brittle failures compartmentalise the system . differently , structures with large plastic capacity tend to collapse entirely after a sequence of base cutting . this result is a consequence of the fact that , in presence of large plastic capacity , we studied frames with thin structural elements . in facts , columns with larger and thus more realistic cross sectional size and reinforcement would not fail because of base cutting . collision - driven mechanisms also determine the outcome of the fragmentation process . we showed that the fragment mass distribution does not depend on the strength and stiffness of the structural elements . it is namely influenced by the plastic capacity of the elements . in structures with large plastic capacity the fragments are more massive and represent an extra cost in controlled demolitions processes . in the present paper , for clearer interpretability of the results , we limited ourselves to simple geometry , collisions , and constitutive models . implementing more sophisticated collision models , e.g. using polyhedral discrete elements , or enabling discrete elements to fragment ( see e.g. @xcite ) , as well as rate effects are future challenges . we also neglected shear failures , since the structural elements were sufficiently small and slender , but this hypothesis should be removed to deal with structures made of large elements . these models should also be refined if the aim is to simulate in detail the collapse of specific real buildings . nevertheless , already at the present state , several interesting studies can be conducted to analyze the response to earthquakes and to investigate the influence of material disorder , geometric uncertainties , overall geometry , and structural connections . experimental validation remains problematic in the field of progressive collapse , because of difficulty in monitoring collapse of complex buildings , and due to problems concerning repeatability of experiments . further discussion about monitoring collapse of buildings for model validation , especially concerning demolitions , can be found in @xcite . 23 [ 1]#1 [ 1]`#1 ` urlstyle [ 1]doi : # 1 z.p . baant and m. verdure . mechanics of progressive collapse : learning from the world trade center and building demolition . mech .- asce _ , 1330 ( 3):0 308319 , 2007 . carmona , f.k . wittel , f. kun , and herrmann h.j . fragmentation processes in impact of spheres . e _ , 770 ( 5):0 243253 , 2008 . cherepanov and i.e. esparragoza . progressive collapse of towers : the resistance effect . _ , 143:0 203206 , 2007 . chiaia and e. masoero . analogies between progressive collapse of structures and fracture of materials . _ , 1540 ( 1 - 2):0 177193 , 2008 . grierson , l. xu , and y. liu . progressive - failure analysis of buildings subjected to abnormal loading . _ comput.aided civ . _ , 20:0 155171 , 2005 . d. hartmann , m. breidt , v. van nguyen , f. stangenberg , s. hhler , k. schweizerhof , s. mattern , g. blankenhorn , b. mller , and m. liebscher . structural collapse simulation under consideration of uncertainty - fundamental concept and results . _ , 860 ( 21 - 22):0 20642078 , 2008 . d. isobe and m. tsuda . seismic collapse analysis of reinforced concrete framed structures using the finite element method . d. _ , 320 ( 13):0 20272046 , 2004 . g. kaewkulchai and e.b . dynamic behavior of planar frames during progressive collapse . in _ proceedings of the 16th asce engineering mechanics conference _ , 2003 . university of washington , seattle , july 16 - 18 , 2003 . g. kaewkulchai and e.b . williamson . modelling the impact of failed members for progressive collapse analysis of frame structures . _ j. perform . asce _ , 200 ( 4):0 375383 , 2006 . k. khandelwal , s. el - tawil , s.k . kunnath , and h.s . macromodel - based simulation of progressive collapse : steel frame structures . _ j. struct . asce _ , 1340 ( 7):0 10701078 , 2008 . m. levy and m. salvadori . _ why buildings fall down ? _ w. w. norton , new york , 1992 . luccioni , r.d . ambrosini , and danesi r.f . analysis of building collapse under blast load . _ , 26:0 6371 , 2004 . m.a . maes , k.e . fritzsons , and glowienka s. structural robustness in the light of risk and consequences analysis . _ _ , 160 ( 2):0 101107 , 2006 . s. marjanishvili and e. agnew . comparison of various procedure for progressive collapse analysis . _ j. perform . _ , 200 ( 4):0 365374 , 2006 . e. masoero , p. vallini , a.p . fantilli , and b.m . energy - based study of structures under accidental damage . _ key engineering materials _ , 417 - 418:0 557560 , 2010 . progressive collapse basics . _ modern steel constr . _ , 440 ( 3):0 3744 , 2004 . t. pschel and t. schwager . _ computational granular dynamics_. springer - verlag gmbh , berlin , 2005 . pretlove , m. ramsden , and a.g . dynamic effects in progressive failure of structures . . j. impact eng . _ , 110 ( 4):0 539546 , 1991 . . progressive collapse of the world trade center . 1340 ( 2):0 125132 , 2008 . val and e.g. val . robustness of framed structures . _ , 160 ( 2):0 108112 , 2006 . _ progressive collapse assessment of tall buildings_. phd thesis , london imperial college , uk , 2007 . vlassis , b.a izzudin , a.y . elghazouli , and d.a . nethercot . design oriented approach for progressive collapse assessment of steel framed buildings . _ , 160 ( 2):0 129136 , 2006 . wittel , f. kun , h.j . herrmann , and b.h . fragmentation of shells . _ , 930 ( 3):0 035504 , 2004 . @xmath190 & @xmath191 & @xmath192 + + specific weight & @xmath193 & kg / m@xmath194 & 2500 + young modulus & @xmath23 & n / m@xmath30 & 30@xmath195 + shear modulus & @xmath24 & n / m@xmath30 & @xmath196 + compressive yield stress & @xmath197 & n / m@xmath30 & 20@xmath198 + ultimate shortening & @xmath199 & - & 0.0035 + + young s modulus & @xmath200 & n / m@xmath30 & 200@xmath195 + yield stress & @xmath201 & n / m@xmath30 & 440@xmath198 + yield strain & @xmath202 & - & 0.0022 + ultimate strain & @xmath203 & - & 0.05 + @xmath190 & @xmath191 & @xmath192 + + elongation & @xmath204 & ns / m & @xmath205 + torsion & @xmath206 & nms & @xmath173 + bending & @xmath207 & nms & @xmath208 + + elongation & @xmath209 & - & @xmath210 + compression & @xmath95 & - & @xmath211 + + columns & @xmath96 & rad & @xmath212 + & & & @xmath213 + beams & slabs & @xmath96 & rad & @xmath212 + & & & @xmath214 + + elongation & @xmath115 & @xmath215 & @xmath216 + compression & @xmath116 & @xmath215 & @xmath217 + rotation & @xmath117 & @xmath218 & @xmath219 + & & & @xmath220 + @xmath190 & @xmath191 & @xmath192 + + sphere - sphere & @xmath9 & n / m@xmath194 & @xmath221 + sphere - ground & @xmath222 & n / m@xmath194 & @xmath223 + + sphere - sphere & @xmath224 & ns / m & @xmath225 + sphere - ground & @xmath226 & ns / m & @xmath227 + + coulomb friction & @xmath228 & ns / m & @xmath229 + dynamic friction & @xmath230 & ns / m & @xmath229 + rolling friction & @xmath231 & nms & @xmath232 + + coulomb friction & @xmath233 & ns / m & @xmath234 + dynamic friction & @xmath235 & ns / m & @xmath234 + rolling friction & @xmath236 & nms & @xmath237 +	in this paper , we study the progressive collapse of 3d framed structures made of reinforced concrete after the sudden loss of a column . the structures are represented by elasto - plastic euler bernoulli beams with elongation - rotation failure threshold . we performed simulations using the discrete element method considering inelastic collisions between the structural elements . the results show what collapse initiation and impact - driven propagation mechanisms are activated in structures with different geometric and mechanical features . namely , we investigate the influence of the cross sectional size and reinforcement @xmath0 and of the plastic capacity @xmath1 of the structural elements . we also study the final collapse extent and the fragment size distribution and their relation to @xmath0 , @xmath1 and to the observed collapse mechanisms . finally , we compare the damage response of structures with symmetric and asymmetric reinforcement in the beams . * keywords : * frames , progressive collapse , robustness , discrete elements
the main conjecture in de broglie s thesis is _ the existence of a periodic phenomenon of an unknown nature , associated to every portion of isolated energy , and related to its proper mass by planck - einstein s equation _ @xcite , so that a particle of mass @xmath5 at rest should have an internal frequency @xmath6 and when the particle moves at a velocity @xmath7 this frequency will be @xmath8 , where @xmath9 is the relativistic time dilation factor . properly speaking , a moving electron has a greater energy @xmath10 and therefore a greater frequency @xmath11 according to the above statement , but if this internal periodic motion has a physical reality will have the frequency proposed by de broglie @xmath8 when measured by an inertial observer who sees the electron moving at the speed @xmath7 , and according to the time dilation of special relativity . using this conjecture , gouanre et al . @xcite tried to get evidence of this electron internal clock frequency by analysing the interaction of a beam of electrons with the atoms of a @xmath12 m thick silicon crystal aligned along the @xmath13 direction . if the momentum of the electron beam is the appropiate , they will interact in a kind of transversal resonant scattering , whenever the distance travelled by the electron during a period @xmath14 will be related to the separation @xmath15 between the atoms of the crystal . when @xmath16 this will produce a lateral force such that the outgoing beam in the forward direction will show a decrease in the number of outgoing electrons for that resonant momentum @xmath17 . this resonant momentum is @xmath18 mev@xmath1 while their experiment shows a value of @xmath19 mev@xmath1 , a 0.28% higher than the predicted theoretical value ( see figure [ fig:1 ] ) . their monte carlo calculation , using parabolic potentials to describe the atomic structures , shows a bad fitting with the experimental result for this internal frequency ( curve ( a ) of figure [ fig:1 ] ) , while a clear matching ( curve ( b ) of figure [ fig:1 ] ) , except for the above 0.28% shift , between the expected and the experimental outcome for a momentum range between @xmath20 mev@xmath1 and @xmath21 mev@xmath1 and when the electron frequency is taken twice de broglie s frequency @xmath22 or @xmath23 . this resonant momentum is @xmath24 mev@xmath1 , it is outside of the experiment range , and the observed peak corresponds to the second harmonic . this frequency corresponds to the zitterbewegung frequency or frequency found by dirac of the internal motion of the electron . transcription of figure 4 of 2008 reference @xcite , which shows the experimental outcome of the detected number of electrons versus the linear momentum @xmath17 of the electron beam in mev@xmath1 ( dotted line ) . curve ( a ) represents their monte carlo calculation for de broglie s frequency @xmath25 . curve ( b ) represents their monte carlo calculation by considering that the electron internal frequency is twice de broglie s frequency @xmath22 . it matches with the experimental result except for a shift from @xmath26 mev@xmath1 to @xmath27 mev@xmath1,width=264 ] we are going to produce this analysis based upon a classical electron model found from a very general formalism for describing spinning particles @xcite . recently @xcite , we have given different arguments justifying that the centre of charge and centre of mass of a spinning particle are different points , and , for the electron at rest , the centre of charge is moving in circles , at the speed of light , around the centre of mass with a frequency @xmath28 , twice as much as de broglie s frequency and coherently matched with dirac s internal frequency . this internal frequency changes as @xmath29 when the electron moves , because of the special relativity time dilation . it is this relative motion of the centre of charge which has a clear interpretation as the schroedinger s zitterbewegung , and therefore a clear internal periodic phenomenon associated to a spinning particle . in the kinematical formalism @xcite , an elementary particle is , by definition , a mechanical system which in addition to being indivisible , as a consequence of the atomic hypothesis @xcite , it can never be deformed so that all allowed states are only kinematical modifications of any one of them . this means that when the state of an elementary particle changes there exists another inertial observer who measures the particle in the same state as before . this means that in a variational approach of classical physics , the initial @xmath30 and final @xmath31 states of the evolution of an elementary particle must be related by a transformation of the kinematical group @xmath32 . therefore , the boundary variables of the variational approach , necessarily span a homogeneous space of the kinematical group of space - time symmetries . when quantizing all classical systems characterized by such homogeneous spaces , their hilbert space of pure states carries a projective unitary irreducible representation of the kinematical group @xcite . it thus satisfies wigner s definition of a quantum elementary particle . in this way , the parameters of the kinematical group become the classical variables we use , as the boundary values of the variational formalism for describing an elementary particle . in the relativistic and non - relativistic approach , these variables are reduced to the ten variables @xmath33 and @xmath34 , interpreted respectively as the time , position of the charge , velocity of the charge and orientation of the particle . in the relativistic case we have three disjoint , maximal homogeneous spaces of the poincar group spanned by these variables with the constraint either @xmath35 , @xmath36 or @xmath37 . it is the manifold with @xmath36 , as suggested by the kinematical analysis of reference @xcite , which leads to dirac s equation when quantizing the system . this kinematical analysis states the possibility that the centre of mass and centre of charge of a spinning particle could be different points . in that case , the centre of charge has necessarily a velocity of constant absolute value , unreachable by any inertial observer , so that only the relativistic approach is selected and the charge moves at the speed @xmath38 . the lagrangian depends on these ten kinematical variables and also on the acceleration of the point @xmath39 and on the angular velocity . for a dirac particle , the charge located at point @xmath39 is moving at the speed of light @xmath36 . the classical expression linear in the energy @xmath40 and in the linear momentum @xmath41 , which gives rise to dirac s equation is @xmath42 where the energy @xmath40 is expressed as the sum of two terms , @xmath43 , or translational energy and the other , which depends on the spin of the system , or rotational energy . the spin comes from the dependence of the lagrangian @xmath44 of both , the acceleration @xmath45 , and the angular velocity @xmath46 , and if we define @xmath47 it takes the form @xmath48 the first part @xmath49 , or _ zitterbewegung _ part , is related to the separation between the centre of charge from the centre of mass and takes into account this relative orbital motion . it quantizes with integer values . the second part @xmath50 is the rotational part of the body frame and quantizes with both integer and half - integer values . the total angular momentum with respect to the origin of observer s frame is @xmath51 so that the spin @xmath52 is the angular momentum of the system with respect to the centre of charge @xmath39 , and not with respect to the centre of mass @xmath53 . by this reason , it is not a conserved quantity for a free particle , but satisfies the dynamical equation @xmath54 this is the same dynamical equation satisfied by dirac s spin operator in the quantum case . when expressed dirac s spin @xmath52 and the centre of mass position @xmath53 in terms of the velocity and acceleration of the charge they take , respectively , the form @xmath55 dirac s spin is always orthogonal to the osculator plane of the trajectory of the charge @xmath39 , in the direction opposite to the binormal for a positive energy particle , and in the opposite direction for the antiparticle . this implies a difference in chirality between matter and antimatter . point @xmath53 is always different from point @xmath39 , because the observable @xmath56 thus justifying the conjecture that for a spinning particle the centre of mass and centre of charge are different points . the acceleration of the charge is pointing from @xmath39 to the centre of mass @xmath53 , as it corresponds to a helix . it is shown that the dynamical equation of point @xmath39 for the free particle and in the centre of mass frame is given by @xmath57 where the spin vector @xmath52 is constant in this frame , as depicted in fig . [ fig:3 ] . the radius of the zitterbewegung motion is @xmath58 , and the angular velocity @xmath59 . when quantizing the system the classical parameter @xmath60 . the frequency of this internal motion and in the centre of mass frame is @xmath28 , i.e. , twice de broglie s frequency , and the radius is @xmath61 , is half compton s wavelength . motion of the charge of the electron at the speed of light in the centre of mass frame . the magnetic moment of the particle is produced by the motion of the charge . the frequency of this internal motion in this frame is @xmath28 . the total spin @xmath52 is half the value of the zitterbewegung part @xmath62 when quantizing the sytem , so that when expressing the magnetic moment in terms of the total spin we get a @xmath63 gyromagnetic ratio @xcite . the body frame attached to the point @xmath39 rotates with angular velocity @xmath64 , has not been depicted.,width=264 ] when seen from an arbitrary observer ( see figure [ fig:4 ] ) , the motion of the charge for a longitudinaly polarized electron is a helix , so that according to ( [ eq : dynspin ] ) dirac s spin precess around the direction of the conserved linear momentum @xmath41 . the spin with respect to the centre of mass is defined as @xmath65 it is a conserved quantity for a free particle . the centre of mass velocity is @xmath66 , and the linear momentum is written as usual as @xmath67 . this means that the transversal motion of the charge is at the velocity @xmath68 . the centre of charge of a moving electron takes a time @xmath69 times longer than for an electron at rest to complete a turn , which is a clear consequence of the time dilation . the faster the centre of mass of the electron moves the slower is the rotation frequency of the centre of charge around the centre of mass . the internal clock of a fast electron is running slower and we are going to take advantage of this fact to get the resonant interaction with the si atoms . precession of dirac s spin @xmath52 along the linear momentum @xmath41 for a moving electron . the tranversal motion of the charge takes a time @xmath69 longer than when the centre of mass is at rest , to complete a turn and , therefore , the frequency of the internal motion of the charge is @xmath29 . the spin with respect to the centre of mass @xmath70 , is a constant of the motion for a free particle.,width=264 ] in figure [ si ] we show the unit cell of a silicon crystal and the separation @xmath71 between neighbouring atoms on the @xmath72 plane along which the electron beam is sent in the gouanre et al . experiment @xcite . in the figure [ fig:2 ] we show two possible free motions of a polarized electron . in part ( a ) we describe the motion of the centre of mass and centre of charge of an electron polarized along the direction of motion @xmath73 . in part ( b ) the electron is polarized perpendicularly to the page along the @xmath74 axis . in both cases the centre of mass motion is in the direction @xmath13 of the crystalographic plane , where the dots represent the location of the si nuclei on that plane . for electrons polarized along the @xmath75 direction , where the centre of charge motion is contained in the plane @xmath76 , the description is equivalent to the last one , where the atoms will be arranged on the @xmath76 plane . the frequency of the internal motion of a moving electron is @xmath77 , so that the distance travelled by the centre of mass during a period is @xmath78 . there will be a resonant scattering whenever @xmath79 or when @xmath80 for any natural number @xmath81 . in the first case , the particle moves faster and there will be a resonant interaction every @xmath82th atom , while in the second case the momentum will be smaller , the particle will interact with each atom but every @xmath83 periods of the electron . the previsible values of these momenta are @xmath84 @xmath85 @xmath86 in the gouanre et al . experiment @xcite the range of the electron beam was prepared between 54 to 110 mev@xmath1 , so that only the resonant frequency for @xmath3 , 80.874 mev@xmath1 was available within that range , although they mention that _ an observation could also be possible at harmonic frequencies , for example at corresponding momenta of 161.748 mev@xmath1 and 40.437 mev@xmath1 , etc . , probably with reduced intensity._. our conjecture is that the momenta of small value @xmath4 are more favorable to produce a larger transversal scattering because the velocity of the beam is smaller , the particle remains in the crystal a longer time and the interaction is produced with every atom , and , therefore , the ratio of the transversal momentum transfer to the longitudinal momentum of the beam is larger , so that a greater number of electrons will be withdrawn from the forward direction , while for the larger momenta @xmath87 the ratio of the transversal perturbation to the longitudinal velocity is smaller and only a smaller number of atoms participate in the resonant scattering . this means that smaller the momenta the larger the depth of the counts of the resonant peak . for very small @xmath4 , and depending on the experimental resolution , it will be difficult to discriminate among the corresponding resonance peaks . for example , for @xmath88 and 101 , the corresponding momenta are @xmath89 which differ from each other of around @xmath90 mev@xmath1 which seems to be smaller than the resolution of the experiment . silicon crystal unit cell , showing the direction @xmath13 on the @xmath91 plane , and the separation @xmath15 between neighbouring atoms on that plane.,width=264 ] part ( a ) describes a free motion of an electron polarized longitudinally along the @xmath73 axis , while part ( b ) describes the free motion of an electron polarized transversally in the direction @xmath74 , perpendicular to the picture and the motion of the centre of charge is contained on the @xmath91 plane . the straight line represents the centre of mass motion and the curly line is the trajectory described by the centre of charge . the range @xmath92 m of the transversal motion of the centre of charge is compton s wave length , and @xmath93 m represents the separation between atoms of the si crystal on the plane corresponding to the @xmath13 direction of the electron beam . both lengths are clearly not represented at the same scale . @xmath44 is the distance travelled by the electron during a period . there will be resonant scattering whenever @xmath79 or @xmath80 , for @xmath81,width=264 ] 99 de broglie l 1924 thse de doctorat . sommaire : _ nous admettons dans le prsent travail lexistence dun phnomne periodique dune nature encore prciser qui serait li tout morceau isol dnergie et qui dpendrait de sa masse propre par lquation de planck - einstein . _ gouanre m , spighel m , cue n , gaillard mj , genre r , kirsch r , poizat jc , remillieux j , catillon p and roussel l 2005 _ annales fond . louis de broglie _ * 30 * 109 + 2008 _ found . phys . _ * 38 * 659 rivas m 2001 _ kinematical theory of spinning particles _ ( dordrecht : kluwer ) + see also some other works related to this formalism through the webpage rivas m 2008 ( _ preprint _ arxiv:0807.2512 ) rivas m 2008 _ j. phys . a : math . * 41 * 304022 ( _ preprint _ arxiv:0709.0192v2 ) rivas m 1994 _ j. math . phys . _ * 35 * 3380 rivas m , aguirregabiria j m and hernndez a 1999 _ phys . a _ * 257 * 21 aguirregabiria jm 2006 _ dynamics solver _ , computer program for solving different kinds of dynamical systems , which is available from his author through the web site at the server of the theoretical physics dept . of the university of the basque country , bilbao , spain	the existence of an internal frequency associated to any elementary particle conjectured by de broglie is compared with a classical description of the electron , where this internal structure corresponds to the motion of the centre of charge around the centre of mass of the particle . this internal motion has a frequency twice de broglie s frequency , which corresponds to the frequency found by dirac when analysing the electron structure . to get evidence of this internal electron clock a kind of experiment as the one performed by gouanre et al . @xcite will show a discrete set of momenta at which a resonant scattering effect , appears . the resonant momenta of the electron beam are given by @xmath0 mev@xmath1 , @xmath2 , where only , the corresponding to @xmath3 , was within the range of gouanre et al . experiment . the extension of the experiment to other values of @xmath4 , would show the existence of this phenomenon .
synchronization processes on complex networks has received a lot of attention during the last decades @xcite . the interplay between the dynamical evolution of oscillators and their local interactions ( as given by the complex topology of a network ) usually results in non - trivial phenomena of relevance to physical , biological , technological and social systems . first introduced by pecora and carroll @xcite , the master stability function ( msf ) is nowadays one of the main theoretical methods for the study of network synchronization . msf is indeed a powerful tool to analyze the stability of the synchronization manifold when identical systems of oscillators are diffusively coupled . originally applied to undirected networks , the msf approach has been later extended to investigate enhancements and optimization of complete synchronization in weighted and asymmetric topologies ( see @xcite , and references therein ) . in @xcite the authors stated the so - called heterogeneity paradox , i.e. the fact that heterogeneous networks , wherein distances between nodes are relatively short , are less stable , in terms of synchronization , than their homogeneous counterparts . soon after , a proper and adequate weighting of the link strengths was shown to overcome this paradox , based again on concepts sparkling from the msf formalism @xcite . following works , have shown how different network s topological features influence the stability of the synchronous state , such as : heterogeneity of the node degree , degree - degree correlations , average shortest - path , betweenness centrality or clustering . these latter studies indicate that altering the structure of a network may result in maximizing the stability of the synchronous state , thus achieving a _ maximally stable synchronization structure _ @xcite . enhancement of the networks synchronizability can also be achieved by the application of genetic algorithms increasing the stability of the synchronized state . in this case , the networks self - organize by disconnecting the hubs and connecting peripheral nodes , thus increasing the homogeneity and leading to what is known as _ entangled networks _ @xcite . in our study , we report the enhancement of the stability of complete synchronization of an ensemble of dynamical units , when coupled simultaneously in different dimensions . we are concerned with a multivariable coupling , where the dynamical systems are coupled through different dimensions according to a certain structure of connections ( see fig.[fig01 ] for a schematic illustration ) . in particular , we consider a generic dynamical system whose associated vector state @xmath0 ( with @xmath1 ) evolves according to @xmath2 . each one of the @xmath3 state variables of the dynamical system at a given node can be coupled to the corresponding variable of any of the other systems ( i.e. , nodes ) of the network . equivalently , we can think of our system as a network with @xmath4 layers , each one accounting for the structure of couplings at each variable of the system . this multilayer point of view illustrated in fig.[fig01 ] is , in fact , just accounting for a multivariable coupling between the nodes of a network , nevertheless it will help us to provide a more concrete representation of the structure of the system , and possible connections to applications . so we will make use of it at certain points . if the coupling between oscillators does not include some of the state variables , _ i.e. _ @xmath5 , the topology of the corresponding layers to those variables would be trivially given by a zero adjacency matrix , so we would not consider them to be proper layers ( as is the case of the layer corresponding to variable @xmath6 in fig.[fig01 ] ) . for simplicity , we consider a bidirectional coupling between the same variables of each system ( i.e. each layer is an undirected network ) . this is illustrated in fig . [ fig01 ] with an example of the case @xmath7 and @xmath8 . interestingly , our framework connects with the so - called _ hypernetwork _ formalism introduced by sorrentino @xcite . in this latter work , the author shows that a msf approach to hypernetwork synchronization is possible when the laplacian matrices @xcite of different layers ( accounting for the coupling through each variable ) have the same basis of eigenvectors , _ i.e. _ , when they are simultaneously diagonalizable . this is a condition that has been shown to be fulfilled for two layers in three cases : ( i ) the laplacian matrices of the different layers are commuting ( a condition that automatically allows for a msf approach whatever the number of layers if the laplacian matrices form a pairwise commuting set ) , ( ii ) one of the two layers is unweighted and fully connected , or ( iii ) one of the two layers has an adjacency matrix of the form @xmath9 with @xmath10 . additionally , ref . @xcite contains an extension of the approach in @xcite to more general topologies by making use of a simultaneous block diagonalization of laplacian matrices corresponding to different layers , thereby decreasing the dimensionality of the linear stability problem . in our work , we consider the topology to be the same in each layer , trivially falling , from the of view of hypernetworks , into category ( i ) of ref . @xcite . this way , we present a study on how the stability of the synchronous state is enhanced by finding an optimal balance for the coupling between the different variables in a network of identical oscillators with multivariable coupling . on the one hand , we provide results based on extensive numerical simulations of networks of rssler - like oscillators to show the applicability of the proposed ideas and how the msf can help us to find the adequate balance between the couplings that optimizes the stability of the synchronous state of a network . on the other hand , by constructing an electronic version of the model , we show that these predictions are in good agreement with the experimental evidences in spite of the idealizations used in the theoretical treatment . in this section we explain how stability of the synchronous state can be enhanced by engineering a multivariable coupling function between nodes in a network and what balance between coupling variables is the most adequate . for the sake of concreteness we focus on a set of rssler - like oscillators @xcite coupled to their neighbors through both the @xmath11 and the @xmath12 variables , whose dynamics evolve according to the following equations : @xmath13 ) \\ \nonumber \label{system1 } \dot{y}_i&=&-\alpha_{2 } ( -\gamma x_i+\left ( 1-\delta \right ) y_i -d_r \sigma_y\phi \sum_{j=1}^na^y_{ij } \left [ y_j - y_i\right ] ) \\ \label{system2 } \dot{z}_i&=&-\alpha_{3}\left ( g_{x_i}+z_i\right)\;,\end{aligned}\ ] ] where @xmath14 , @xmath15 , @xmath16 , @xmath17 , @xmath18 , @xmath19 , @xmath20 , @xmath21 , @xmath22 , @xmath23 and @xmath24 and @xmath25 account for the coupling strengths of variables @xmath11 and @xmath12 . as we explain below , this chaotic oscillator has the highly non - trivial characteristic of being quite robust when implemented in electronic circuits . the adjacency matrices @xmath26 and @xmath27 contain the topology of each of two layers , each one accounting for the coupling through the @xmath11 and @xmath11 variables . elements @xmath28 and @xmath29 are one when nodes @xmath30 and @xmath31 are connected and zero otherwise . with these parameters the oscillators display chaotic dynamics due to the nonlinearity introduced in @xmath32 , which consists on a piecewise function defined as : @xmath33 the coupling between oscillators is here controlled by two parameters : @xmath34 being the coupling strength and @xmath35 controlling how the coupling strength is distributed between variables @xmath11 and @xmath12 . this way , @xmath36 ( @xmath37 ) leads to a coupling restricted to variable @xmath11 ( @xmath12 ) , while a sweep of @xmath35 in the interval @xmath38 $ ] allows for a weighted combination of both @xmath11 and @xmath12 variables . notice that the role of the parameter @xmath35 is therefore that of exploring the consequences of unevenly distributed coupling on the stability of synchronization , which in the past were the object of specific studies in space extended systems @xcite and weighted monolayer graphs @xcite . here however , the difference in the weight assigned to each variable introduced by @xmath35 implies a different balance of two system s variables in the coupling configuration , but does not affect the un - directionality of each one of the network s links . we now apply the msf formalism to study how the amount of coupling effectively distributed among the two coupling variables affects the stability of the synchronous state of the network . denoting the coupling functions of each variable as @xmath39 he dynamics of each node is then given by : @xmath40 + \ d_r \sigma_y \displaystyle\sum_{j=1}^n \mathcal{a}^y_{ij}\left[\ , \textbf{h}_y(\mathbf{x}_j ) -\textbf{h}_y(\mathbf{x}_i ) \right]\\ \nonumber & = & \textbf{f}(\mathbf{x}_i)\ + \ ( 1-d_r ) \sigma_x \displaystyle\sum_{j=1}^n \left(\mathcal{a}^x_{ij } - \delta_{ij } \left[\displaystyle\sum_{j=1}^n\mathcal{a}^x_{ij}\right]\right ) \textbf{h}_x(\mathbf{x}_j ) + \ \sigma_y \displaystyle\sum_{j=1}^n \left(\mathcal{a}^y_{ij } - \delta_{ij } \left[\displaystyle\sum_{j=1}^n\mathcal{a}^y_{ij}\right]\right ) \textbf{h}_y(\mathbf{x}_j)\\ \nonumber & = & \textbf{f}(\mathbf{x}_i)\ -\ ( 1-d_r ) \sigma_x \displaystyle\sum_{j=1}^n \mathcal{l}^x_{ij } \textbf{h}_x(\mathbf{x}_j ) -\ d_r \sigma_y \displaystyle\sum_{j=1}^n \mathcal{l}^y_{ij } \textbf{h}_y(\mathbf{x}_j)\\ \label{fullsys}\end{aligned}\ ] ] where @xmath41 stands for the kronecker delta , and @xmath42 are the laplacian matrices @xcite describing the coupling through variables @xmath11 and @xmath12 respectively . if we consider that @xmath43 and we restrict our analysis to the case of identical coupling topologies for all variables of the system , i.e. @xmath44 , and , in turn , @xmath45 , eq . [ fullsys ] reads : @xmath46 where the coupling function is @xmath47 . this way , eq . [ syssimpl ] is basically the classical equation describing the evolution of a diffusively coupled systems , with the particularity that the coupling function @xmath48 depends on the parameter @xmath35 accounting for how the total amount of coupling is divided between the coupling variables of the system . the dependence on @xmath35 leads to a _ msf that describes the stability of the synchronization manifold . varying the value of @xmath35 we obtain a family of msfs that allows to evaluate how the stability of the synchronized manifold is affected by shifting the coupling from one layer to the other . the independent variable @xmath49 of the msf is related with the eigenvalues of the laplacian matrix ( @xmath50 ) and the synchronization manifold will be stable when , for all eigenvalues of the laplacian matrix @xmath51 , the corresponding @xmath49 leads to a value of the msf that is negative ( for a given value of @xmath34 ) . taking into account that @xmath52 ( since the laplacian is a zero row sum matrix ) and that @xmath53 , dynamical systems can be classified , depending on the shape of its corresponding msf . the classification includes : _ i ) _ class i systems , whose msf is always positive for any value of @xmath49 so that the system can not be synchronized for any topology or coupling strength , _ ii ) _ class ii systems , when the msf is positive for low values of @xmath49 and becomes negative when a threshold value @xmath54 is achieved ( being the synchronization manifold stable when @xmath55 ) and _ iii ) _ class iii , when the msf has two zeros at @xmath54 and @xmath56 , leading to a stable synchronized manifold when @xmath57 and @xmath58 simultaneously . ( color online ) master stability function msf(@xmath49 ) as a function of the coupling fraction @xmath35 between the variables to be coupled . in ( a ) , solid and dashed black lines are , respectively , the msf for a coupling through variables @xmath11 and @xmath12 exclusively . colored lines show the msf obtained for different values of @xmath35 , as indicated in the legend . in ( b ) , msfs for three different values of the parameter @xmath49 : @xmath59 , @xmath60 and @xmath61 . note that , for a sizable range of @xmath35 , the multivariable coupling leads to lower values of the msf , indicating a region where the stability against external perturbations is higher.,title="fig : " ] ( color online ) master stability function msf(@xmath49 ) as a function of the coupling fraction @xmath35 between the variables to be coupled . in ( a ) , solid and dashed black lines are , respectively , the msf for a coupling through variables @xmath11 and @xmath12 exclusively . colored lines show the msf obtained for different values of @xmath35 , as indicated in the legend . in ( b ) , msfs for three different values of the parameter @xmath49 : @xmath59 , @xmath60 and @xmath61 . note that , for a sizable range of @xmath35 , the multivariable coupling leads to lower values of the msf , indicating a region where the stability against external perturbations is higher.,title="fig : " ] now we investigate the synchronization properties of the system depending on the distribution of the coupling strengths among the coupling variables . in fig . [ fig02](a ) we show the msf obtained for the rssler - like system coupled through the @xmath11 variable ( @xmath36 , black solid line ) , the @xmath12 variable ( @xmath62 , black dashed line ) and simultaneously the @xmath11 and @xmath12 variables ( colored lines ) . a range of @xmath35 values is swept in order to show the gradual changes in the stability of the synchronous state as the coupling of one or the other variable is enhanced . as expected , coupling introduced only through variable @xmath11 leads the system to be class iii while it becomes class ii when coupled only through the @xmath12 variable @xcite . interestingly , when the coupling is introduced simultaneously through the two variables we obtain a msf that is not a linear combination of the isolated layers . in particular , the sweep of @xmath35 leads to a family of msfs that are class ii ( at least for the ten cases plotted in fig . [ fig02](a ) ) , thus synchronizing the network when the condition @xmath63 is fulfilled , which is in principle achievable for any topology by just using a sufficiently high @xmath34 . figure [ fig02](a ) also points out an important consequence , namely that using a multivariable coupling ( while maintaining the overall coupling strength ) leads to values of the msf that are more negative than those obtained when coupling the systems through a unique variable , indicating that the stability of the synchronized manifold is higher for multivariable couplings . this is somehow to be expected , since a simple change of coordinates would lead the dynamics of the rssler system to be described as a combination of the actual @xmath11 , @xmath12 and @xmath6 variables . however , by studying how the msf modifies its shape as a function of the combination of the variables of the system we can find the regions with the highest stability . this fact can be observed in fig . [ fig02](b ) , where the msf is plotted as a function of @xmath35 for three different values of @xmath49 . we can observe a minimum of the msf at intermediate values of @xmath35 for @xmath60 and @xmath61 , _ i.e. _ , a @xmath35 value where the stability of the synchronized manifold is the largest . this result is also observed within a range of values of @xmath49 ( not shown here ) . ( color online ) robustness of the synchronized state . synchronization error @xmath64 as a function of the noise strength @xmath65 and the coupling ( rescaled as @xmath66 ) . panels correspond to : ( a ) @xmath36 , ( b ) @xmath37 and ( c ) @xmath67 . the network coupled through both the @xmath11 and @xmath12 variables ( @xmath67 ) is the one showing better performance against noise perturbations . the synchronization error in the system is computed as @xmath68 , where the normalizing factor corresponds to the total number of oscillator pairs in the network . left plots correspond to the msf obtained for each value of @xmath35.,title="fig:",scaledwidth=44.0% ] ( color online ) robustness of the synchronized state . synchronization error @xmath64 as a function of the noise strength @xmath65 and the coupling ( rescaled as @xmath66 ) . panels correspond to : ( a ) @xmath36 , ( b ) @xmath37 and ( c ) @xmath67 . the network coupled through both the @xmath11 and @xmath12 variables ( @xmath67 ) is the one showing better performance against noise perturbations . the synchronization error in the system is computed as @xmath68 , where the normalizing factor corresponds to the total number of oscillator pairs in the network . left plots correspond to the msf obtained for each value of @xmath35.,title="fig:",scaledwidth=44.0% ] ( color online ) robustness of the synchronized state . synchronization error @xmath64 as a function of the noise strength @xmath65 and the coupling ( rescaled as @xmath66 ) . panels correspond to : ( a ) @xmath36 , ( b ) @xmath37 and ( c ) @xmath67 . the network coupled through both the @xmath11 and @xmath12 variables ( @xmath67 ) is the one showing better performance against noise perturbations . the synchronization error in the system is computed as @xmath68 , where the normalizing factor corresponds to the total number of oscillator pairs in the network . left plots correspond to the msf obtained for each value of @xmath35.,title="fig:",scaledwidth=44.0% ] with the aim of testing the predictions of the msf , we simulate a network of @xmath69 bidirectionally coupled rssler - like systems , see fig . [ fig01 ] and eq . ( [ system ] ) . we introduce an additive gaussian noise term ( with zero mean ) in the @xmath11 variable whose strength is given by the parameter @xmath65 . next , we calculate the synchronization error @xmath64 of the network as a function of the coupling @xmath34 and the noise strength @xmath65 , for different values of @xmath35 . this way , we are able to check the robustness of the stability of the synchronized manifold depending on how the coupling is distributed among the @xmath11 and @xmath12 variables of the system . to facilitate the comparison with the msf predictions , we plot the results as a function of @xmath70 , instead of @xmath34 , being @xmath71 for the particular network structure shown in fig . [ fig01 ] . this choice can be justified on the grounds that all the important information from the point of view of the synchronization is given by the msf value of the eigenmode associated to @xmath72 , as the eigenmode associated to @xmath73 is never pushed beyond the boundaries of the synchronization region within the experimentally accessible range of coupling strengths . figure [ fig03 ] shows the two - dimensional plots @xmath74 obtained for @xmath36 ( coupling only through the @xmath11 variable ) , @xmath67 ( @xmath11 and @xmath12 couplings are equally active ) and @xmath37 ( coupling only through the @xmath12 variable ) . for each of these three cases , the coupling strength @xmath34 is modified and the synchronization error between all oscillators is calculated ( see caption of fig.[fig03 ] for details ) . we consider that the network is out of synchrony for synchronization errors @xmath75 ( red regions in fig . [ fig03 ] ) . when the coupling is introduced through the @xmath11 variable ( fig . [ fig03].(a ) ) , the rssler - like oscillators behave as a class iii for any value of the noise strength ( within the range @xmath76 ) , i.e. the system only synchronizes for intermediate values of the coupling strength @xmath49 . when the noise strength @xmath65 is increased , the synchronization error increases , leading to a complete loss of the synchronization for large values of @xmath65 . for couplings such that @xmath77 or larger , the network becomes unstable in the sense that the strong coupling makes the individual oscillators abandon the basin of the attractor , and their dynamics blow up . this type of instability is to be expected , as a similar phenomenon is observed in individual rssler - like oscillators for some initial conditions ( see , e.g. , ref . @xcite ) . the reader should notice that no such problems beset the computation of the msf , as the maximum lyapunov exponent transverse to the synchronization manifold is computed from effectively one individual oscillator along a trajectory that follows the attractor dynamics . thus , there is no contradiction in the fact that the msf determines the attractor dynamics to be synchronizable , while the actual simulation of a network of 6 oscillators never attains attractor dynamics . nevertheless , this is an example of the importance of checking the practical applications of the msf predictions , specially for high values of coupling strengths . when coupling is only introduced through the @xmath78 variable ( fig . [ fig03].(b ) ) the oscillators behave as class ii systems for low values of @xmath65 , as expected . nevertheless , for moderate values of the noise strength ( @xmath79 ) the system shifts to a behavior similar to that of class iii systems , synchronizing only for intermediate values of the coupling strength . finally , when the noise is further increased ( @xmath80 , not shown here ) , the network is not able to reach the synchronized state for any value of the coupling , behaving as a class i system . this way , despite being a class ii system ( when coupled through the y variable ) , the addition of noise can make system behave differently . finally , it is also worth analyzing how the combination of both layers increases the stability of the synchronized manifold . figure [ fig03].(c ) shows the synchronization error for @xmath67 , _ i.e. _ when the coupling is equally distributed among both x and y layers . we can observe that , as suggested by the theoretical predictions shown in fig . [ fig02 ] , the combined coupling of the x and y variables enhances the network stability , so that the synchronization of the system is maintained even for high values of the noise strength . obviously , only for small values of the coupling strength , here measured as @xmath49 , unsynchronized motion is observed , as it is expected in class ii systems . note that the boundaries of the region where synchronization becomes robust show an excellent agreement with the zero crossings of the msf . on the other hand , in the areas where synchronization is most robust against the presence of noise , the optimal @xmath49 from that point of view is not always close to the minimum of the msf . this is not surprising , considering that we are using noise strengths of the order @xmath81 , which are certainly beyond any reasonable definition of _ infinitesimal perturbations _ , a necessary requirement for obtaining the msf . the addition of finite perturbations , being a largely unexplored issue , are beyond of the scope of the msf framework , and require specific numerical studies adapted to each particular topology , even if some general behaviour has been identified @xcite . ( color online ) experimental results . synchronization errors as a function of the noise strength @xmath65 and the coupling ( normalized to @xmath82 ) . as in fig . [ fig03 ] , panels correspond to : ( a ) @xmath36 , ( b ) @xmath37 and ( c ) @xmath67 . region colored in red indicates where the combination of noise and coupling strength leads to a loss of the complete synchronization . the network combining the coupling through the two variables @xmath11 and @xmath12 ( @xmath67 ) is the one showing better performance against noise perturbations . left plots correspond to the msf obtained for each value of @xmath35.,title="fig:",scaledwidth=45.0% ] ( color online ) experimental results . synchronization errors as a function of the noise strength @xmath65 and the coupling ( normalized to @xmath82 ) . as in fig . [ fig03 ] , panels correspond to : ( a ) @xmath36 , ( b ) @xmath37 and ( c ) @xmath67 . region colored in red indicates where the combination of noise and coupling strength leads to a loss of the complete synchronization . the network combining the coupling through the two variables @xmath11 and @xmath12 ( @xmath67 ) is the one showing better performance against noise perturbations . left plots correspond to the msf obtained for each value of @xmath35.,title="fig:",scaledwidth=45.0% ] ( color online ) experimental results . synchronization errors as a function of the noise strength @xmath65 and the coupling ( normalized to @xmath82 ) . as in fig . [ fig03 ] , panels correspond to : ( a ) @xmath36 , ( b ) @xmath37 and ( c ) @xmath67 . region colored in red indicates where the combination of noise and coupling strength leads to a loss of the complete synchronization . the network combining the coupling through the two variables @xmath11 and @xmath12 ( @xmath67 ) is the one showing better performance against noise perturbations . left plots correspond to the msf obtained for each value of @xmath35.,title="fig:",scaledwidth=45.0% ] to test the robustness of the theoretical predictions given by the msf and the numerical simulations , we have implemented the electronic version of the rssler system described in eq . [ system ] . a schematic representation of the experimental design is shown in fig . [ fig04 ] . it consists of an electronic array ( ea ) , a personal computer ( pc ) and a multifunction data card ( daq ) composed by 12 analog to digital converters ( adcs ) and 6 digital to analog converter ( dacs ) . the adcs are used for sampling the variable @xmath11 of the oscillators , while the dacs generate six noisy signals that perturb the dynamics of each node separately . the ea comprises 6 rssler - like electronic circuits forming a spiderweb network ( blue nodes ) , with one central node and five peripheral nodes . each node has two individual electronic couplers , one for the @xmath11 variable and the second for the @xmath12 variable , both controlled by a two digital potentiometers ( xdcp ) , which are adjusted by a digital signal coming from ports p0.0 - 1 ( do ) . p0.0 is used to increase or decrease the resistance of the voltage divisor ( @xmath34 ) , and p0.1 sets the value of the resistance ( the resolution allowing for 100 discretized steps ) . the noisy signals are designed in labview , using the library gaussian white noise vi @xcite that generates six different gaussian - distributed pseudorandom sequences bounded between @xmath83 $ ] . these signals are digitally filtered by a third - order lowpass butterworth filter @xcite with a cutoff frequency of @xmath84 khz . all the experimental process is controlled from a virtual interface developed in labview 8.5 . the experimental procedure is the following : first , @xmath34 and @xmath65 are set to zero and then we introduce the six noisy signals and apply the factor gain ( @xmath65 ) . after a waiting time of 500 ms ( roughly corresponding to @xmath85 cycles of the autonomous systems ) , the signals corresponding to the variables of the 6 circuits are acquired by the analog ports ( ai 0 ; ai 1 ; ... ; ai 5 ) and the synchronization error is calculated and stored in the pc . noisy signals are injected by the digital converters ( ao 0 ; ao 1 ; ... ; ao 5 ) and this part of the process is repeated 100 times ( until the maximum value of @xmath34 is reached ) . finally , @xmath65 is increased to the next value and @xmath34 is swept again . the whole process is repeated 100 times until the maximum value of @xmath65 is reached . figure [ fig05 ] shows the experimental results for a configuration identical to that of the numerical simulations shown in fig . [ fig03 ] . we observe that the qualitative agreement between numerics and experiment is excellent , in spite of unavoidable parameter mismatches in the experimental realization due to the tolerance of the electronic components ( between 5@xmath86 and 10@xmath86 ) . the parameter mismatch , together with the experimental noise , make the oscillators in the network not only slightly different from their mathematical definition , but also non - identical to one another . this way , we confirm experimentally the feasibility of using the msf for evaluating how the coupling through multiple variables enhances the stability of the synchronous state of a network under realistic conditions . we have seen how an adequate distribution of the coupling strength between the variables of a dynamical system leads to an enhancement of the stability of the synchronized manifold . in particular , we have shown that it is possible obtain a msf that depends on the parameter @xmath35 accounting for the distribution of strength , while maintaining the global coupling constant . interestingly , we report the existence of an optimal value of @xmath35 indicating what is the most adequate amount of coupling to be considered at each coupling variable . the optimal value of @xmath35 is independent of the topology of the network , as long as we use the same coupling structure among all variables . using electronic circuits , we have also checked the robustness of the results when noise and parameter mismatch are considered , which confirms the theoretical predictions given by the parametric msf and , in its turn , reveal that the requirement of the oscillators to be identical can be relaxed . the proposed framework of decomposing the different dimensions of the system ( variables ) in interconnected layers paves the way to the use of the multilayer networks tools @xcite to further analyze synchronization phenomena in multivariable coupled systems . indeed , the current theoretical efforts in network theory to define and study complex structures resulting from the interaction of networks , _ e.g. _ interdependent networks and multiplex networks among others , have made great progress in recent times , in showing new emergent phenomena with no counterpart in single ( monolayer ) complex networks @xcite . new developments in that direction could be further extend the results we here present , either using , e.g. , the insight developed within the hypernetwork formalism @xcite , or using new approaches for analyzing multilayer networks @xcite . indeed , our methodology has some limitations that must be further explored in the future . first of all , the fact that the coupling must have the same topological structure at all variables is a strict constraint , since real systems may have different configurations depending on the coupling variable . more general , fully multilayer topologies could be considered by resorting to the hypernetwork formalism introduced in @xcite , and for greater generality one can use the method in ref . @xcite . with this methodology , the case of @xmath87 could be addressed at the cost of introducing some more complexity to the problem . nevertheless , it would be of great interest , since it would raise new questions such as what the adequate combination of topologies would be given a specific distribution of weights @xmath35 . second , since the parametric msf depends on the dynamical system implemented in the network , we can not guarantee the existence of an optimal balance of the distribution of coupling between layers in other dynamical systems , at least until their corresponding msf have been analyzed . authors acknowledge d. papo and p.l . del barrio for fruitful conversations . support from mineco through projects fis2011 - 25167 , fis2012 - 38266 and fis2013 - 41057-p is also acknowledged . aa and jgg acknowledge support from the ec fet - proactive project plexmath ( grant 317614 ) and multiplex ( grant 317532 ) . jgg acknowledges support from mineco through the ramn y cajal program , the comunidad de aragn ( grupo fenol ) and the brazilian cnpq through the pve project of the ciencia sem fronteiras program . aa acknowledges icrea academia and the james s. mcdonnell foundation . r.s.e . acknowledges universidad de guadalajara , culagos ( mexico ) for financial support ( op / pifi-2013 - 14msu0010z-17 - 04 , proinpep - rg/005/2014 , udg - conacyt / i010/163/2014 ) and conacyt ( becas mixtas mzo2015/290842 ) . a network with a connectivity matrix @xmath88 ( @xmath89 if nodes @xmath30 and @xmath31 are connected , and 0 otherwise ) can be represented by the laplacian matrix @xmath90 , where @xmath91 ( @xmath41 denoting the kronecker delta )	synchronization processes in populations of identical networked oscillators are in the focus of intense studies in physical , biological , technological and social systems . here we analyze the stability of the synchronization of a network of oscillators coupled through different variables . under the assumption of an equal topology of connections for all variables , the master stability function formalism allows assessing and quantifying the stability properties of the synchronization manifold when the coupling is transferred from one variable to another . we report on the existence of an optimal coupling transference that maximizes the stability of the synchronous state in a network of rssler - like oscillators . finally , we design an experimental implementation ( using nonlinear electronic circuits ) which grounds the robustness of the theoretical predictions against parameter mismatches , as well as against intrinsic noise of the system .
we consider applications of numerical analytical technique based on modification of our variational - wavelet approach to nonlinear collective models of beam / plasma physics , e.g. some forms of vlasov / boltzmann - like reductions from general bbgky hierarchy ( section 2 ) . these equations are related to the modeling of propagation of intense charged particle beams in high - intensity accelerators and transport systems [ 1 ] , [ 2 ] . in our approach we use fast convergent multiscale variational - wavelet representations , which allows to consider polynomial and rational type of nonlinearities [ 3]-[16 ] , [ 17 ] . the solutions are represented via the multiscale decomposition in nonlinear high - localized eigenmodes ( generalized gluckstern modes , in some sense ) , which corresponds to the full multiresolution expansion in all underlying hidden time / space or phase space scales . in contrast with different approaches we do nt use perturbation technique or linearization procedures . in section 3 after formulation of key points we consider another variational approach based on ideas of para - products and nonlinear approximation in multiresolution approach , which gives the possibility for computations in each scale separately [ 18 ] . we consider representation ( 4 ) below , where each term corresponds to the contribution from the scale @xmath0 in the full underlying multiresolution decomposition as multiscale generalization of old ( nonlinear ) @xmath1 approach [ 1 ] . as a result , fast scalar / parallel modeling demonstrates appearance of high - localized coherent structures ( waveletons ) and pattern formation in systems with complex collective behaviour . let m be the phase space of ensemble of n particles ( @xmath2 ) with coordinates @xmath3 with distribution function @xmath4 and @xmath5 be the n - particle distribution functions ( @xmath6 is permutation group of n elements ) . for s=1,2 we have from general bbgky hierarchy : @xmath7 @xmath8 where partial liouvillean operators are described in [ 17 ] . we are interested in the cases when @xmath9 where @xmath10 are correlation patterns , really have additional reductions as in case of vlasov - like systems . then we have in ( 2 ) , ( 3 ) polynomial type of nonlinearities ( more exactly , multilinearities ) . our goal is the demonstration of advantages of the following representation @xmath11 for the full exact solution for the systems related to equations ( 2 ) , ( 3 ) . it is possible to consider ( 4 ) as multiscale generalization of old ( nonlinear ) @xmath1 approach [ 1 ] . in ( 4 ) each @xmath12 term corresponds to the contribution from the scale @xmath0 in the full underlying multiresolution decomposition @xmath13 of the proper function space ( @xmath14 , hilbert , sobolev , etc ) to which @xmath15 is really belong . it should be noted that ( 4 ) does nt based neither on perturbations nor on linearization procedures . although usually physicists , who prefered computer modelling as a main tool of understanding of physical reality , do nt think about underlying functional spaces , but many concrete features of complicated complex dynamics are really related not only to concrete form / class of operators / equations but also depend on the proper choice of function spaces where operators actully act . moreover , we have for arbitrary @xmath16 in the finite n - mode approximation @xmath17 the following more useful decompositions : @xmath18 in case when @xmath19 is the coarsest scale of resolution and where @xmath20 and bases in scale spaces @xmath21 are generated from base functions @xmath22 by action of affine group of translations and dilations ( the so called `` wavelet microscope '' ) . the following constructions based on variational approach provide the best possible fast convergence properties in the sense of combined norm @xmath23 introduced in [ 17 ] . our five basic points after functional space choice are : 1 . ansatz - oriented choice of the ( multidimensional ) bases related to some polynomial tensor algebra . some example related to general bbgky hierarchy is considered in [ 17 ] . 2 . the choice of proper variational principle . a few projection/ galerkin - like principles for constructing ( weak ) solutions are considered in [ 3 ] - [ 16 ] . it should be noted advantages of formulations related to biorthogonal ( wavelet ) decomposition . the choice of bases functions in scale spaces @xmath24 from wavelet zoo . they correspond to high - localized ( nonlinear ) oscillations / excitations , coherent ( nonlinear ) resonances , etc . besides fast convergence properties of the corresponding variational - wavelet expansions it should be noted minimal complexity of all underlying calculations , especially in case of choice of wavelet packets which minimize shannon entropy . operators representations providing maximum sparse representations for arbitrary ( pseudo ) differential/ integral operators @xmath25 , etc [ 17 ] . 5 . ( multi)linearization . besides variation approach we consider now a different method to deal with ( polynomial ) nonlinearities . we modify the scheme of our variational approach in such a way in which we consider different scales of multiresolution decomposition ( 5 ) separately . for this reason we need to compute errors of approximations . the main problems come of course from nonlinear ( polynomial ) terms . we follow according to the multilinearization ( in case below bilinearization ) approach of beylkin , meyer etc from [ 18 ] . let @xmath26 be projection operators on the subspaces @xmath27 ( 5 ) : @xmath28 and @xmath29 are projection operators on the subspaces @xmath24 : @xmath30 . so , for @xmath31 we have @xmath32 and @xmath33 . it is obviously that we can represent @xmath34 in the following form : @xmath35 in this formula there is no interaction between different scales . we may consider each term of ( [ eq : form1 ] ) as a bilinear mappings : @xmath36 @xmath37 for numerical purposes we need formula ( [ eq : form1 ] ) with a finite number of scales , but when we consider limits @xmath38 we have @xmath39 which is para - product of bony , coifman and meyer [ 18 ] . now we need to expand ( [ eq : form1 ] ) into the wavelet bases . to expand each term in ( [ eq : form1 ] ) we need to consider the integrals of the products of the basis functions ( 7 ) , e.g. @xmath40 where @xmath41 and @xmath42 are the basis functions ( 7 ) . for compactly supported wavelets @xmath43 where @xmath44 depends on the overlap of the supports of the basis functions and @xmath45 as the distance between scales such that for a given @xmath46 all the coefficients in ( [ eq : form4 ] ) with labels @xmath47 , @xmath48 have absolute values less than @xmath46 . for the purposes of computing with accuracy @xmath46 we replace the mappings in ( [ eq : form2 ] ) , ( [ eq : form3 ] ) by @xmath49 @xmath50 since @xmath51 we may consider bilinear mappings ( [ eq : z1 ] ) , ( [ eq : z2 ] ) on @xmath52 . for the evaluation of ( [ eq : z1 ] ) , ( [ eq : z2 ] ) as mappings @xmath53 we need significantly fewer coefficients than for mappings ( [ eq : z1 ] ) , ( [ eq : z2 ] ) . it is enough to consider only coefficients @xmath54 where @xmath55 is scale function . also we have @xmath56 where @xmath57 @xmath58 satisfy the standard system of linear equations and after its solution we can recover all bilinear quantities ( 14 ) . then we apply some variation approach from [ 3]-[16 ] , but in each scale separately . so , after application of points 1 - 5 above , we arrive to explicit numerical - analytical realization of representations ( 4 ) or ( 6 ) . fig.1 demonstrates the first contribution to the full solution ( 6 ) while fig.2 presents ( stable ) pattern as solution of system ( 2)-(3 ) . we evaluate accuracy of calculations according to norm introduced in [ 17 ] . a.n . fedorova and m.g . zeitlin , proc . pac99 , 1614 , 1617 , 1620 , 2900 , 2903 , 2906 , 2909 , 2912 , aps / ieee , new york , 1999 . los alamos preprints : physics/9904039 , 9904040 , + 9904041 , 9904042 , 9904043 , 9904045 , 9904046 , 9904047 . fedorova , m.g . zeitlin , proc . pac2001 , chicago , 790 - 1792 , 1805 - 1807 , 1808 - 1810 , 1811 - 1813 , 1814 - 1816 , 2982 - 2984 , 3006 - 3008 , ieee , 2002 or arxiv preprints : physics/0106022 , 0106010 , 0106009 , 0106008 , 0106007 , 0106006 , 0106005 .	we consider an application of modification of our variational - wavelet approach to some nonlinear collective model of beam / plasma physics : vlasov / boltzmann - like reduction from general bbgky hierachy related to modeling of propagation of intense charged particle beams in high - intensity accelerators and transport systems . we use fast convergent multiscale variational - wavelet representations for solutions which allow to consider polynomial and rational type of nonlinearities . the solutions are represented via the multiscale decomposition in nonlinear high - localized eigenmodes ( waveletons ) . in contrast with different approaches we do not use perturbation technique or linearization procedures . p130 mm * fast calculation methods in * + * collective dynamical models * + * of beam / plasma physics * + * antonina n. fedorova , michael g. zeitlin * + * _ ipme ras , st . petersburg , v.o . bolshoj pr . , 61 , 199178 , russia _ * + * _ e - mail : [email protected]_ * + * _ http://www.ipme.ru/zeitlin.html _ * + * _ http://www.ipme.nw.ru/zeitlin.html _ * presented at the eighth european particle accelerator conference + epac02 + paris , france , june 3 - 7 , 2002
quantum computation promises enormous technological advances in the field of information processing and the quest for its realization has attracted many strong contenders in the field of physics and engineering . this paper is concerned with a scheme for quantum computation put forward by barnes , shilton and robinson,@xcite which falls into the semiconductor quantum dot category.@xcite the proposal for quantum computation is based on the results of ongoing experiments that have demonstrated the capture and transport of single electrons in moving quantum dots . the dots are formed when a surface acoustic wave ( saw ) travels along the surface of a piezoelectric semiconductor containing a two - dimensional electron gas ( 2deg ) . see fig . [ sawdevice ] for a schematic diagram of the device . when the saw is made to pass through a constriction in the form of a quasi - one - dimensional channel ( q1dc ) , the induced piezoelectric potential drags electrons into and along the q1dc . in certain parameter regimes the device transports one electron per potential minimum of the saw.@xcite the spin on the trapped electron represents the physical qubit . quantum computation involves performing qubit operations on the trapped electrons as they move with the speed of the saw . many schemes for quantum computation , such as conventional quantum dots , doped silicon,@xcite superconducting boxes,@xcite and ion traps,@xcite involve static qubits . the surface acoustic wave quantum computer on the other hand is of the `` flying qubit '' type , which include linear optics schemes,@xcite some ion trap schemes with ion shuttling@xcite and schemes based on coherent electron transport in quantum wires.@xcite all these have in common that the carriers of quantum information physically move through space during the computation . flying qubits have the advantage of being able to distribute information quickly over large distances across the quantum circuit when decoherence times are short and to interface with quantum memory registers at fixed locations . another advantage of the saw quantum computation scheme is its ensemble nature . it intrinsically performs time - ensemble computation , in much the same way nmr quantum computation performs molecular - ensemble computations.@xcite time - ensemble computation alleviates the demand for single - shot spin measurements and has the advantage of being robust against small random errors . , width=325 ] this paper considers proposals@xcite for implementing quantum gates on single saw qubits using only static magnetic fields generated by surface magnetic gates . detailed modeling of the gate operation has been accomplished by means of both analytic solutions and numerical simulations of the pauli equation . we show how electrostatically confined moving electrons behave under the influence of various magnetic fields and discuss the implications for quantum computing with surface acoustic wave electrons . the physics of the saw - guided qubit is explained in more detail in section [ section : qubit ] . in sections [ section : unitary ] and [ section : readout ] , we present results on single - qubit unitary gates and single - qubit readout gates based on the stern - gerlach effect . section [ section : neglect ] discusses some of the decoherence processes involved in quantum - dot based schemes . section [ section : summary ] is a summary of the results with parameter regimes of interest for future experiments . we begin by summarizing the quantum computation scheme proposed by barnes , shilton and robinson.@xcite figure [ sawdevice ] shows a schematic diagram of the experimental setup originally designed to demonstrate quantized currents in semiconductors . a nicr / al interdigitated transducer is patterned on a gaas / algaas heterostructure . a narrow depleted q1dc splits the 2deg into two regions , the source and drain . when a high frequency ac signal is applied to the transducer , a saw propagates through the 2deg , producing a periodic piezoelectric potential across the 2deg . the potential drags electrons in the source region through the narrow q1dc constriction into the drain . it has been shown experimentally that over a range of saw power and gate voltages , the current passing through the q1dc is quantized in units of @xmath0 , where @xmath1 is the electronic charge and @xmath2 is the frequency of the saw.@xcite the smallest quantized current observed corresponds to the transport of a single electron in each saw minimum . typically , the saw in gaas moves at @xmath3 at a frequency of around @xmath4 , with an applied power of @xmath5.@xcite these parameters produce currents in the range of nanoamps . given the ability to trap single electrons in the saw minima , the scheme for quantum computation is as follows . it is possible for an array of @xmath6 q1dcs in parallel to capture @xmath6 qubits in every @xmath7th minimum , with a single electron in each q1dc , producing a qubit register along the saw wavefront . @xmath7 can be chosen sufficiently large to ensure that the coulomb interaction between successive qubit registers do not interfere with each separate computation . the qubits move with the minima of the saw , passing through a sequence of static one- and two - qubit gates before arriving at an array of spin readout devices . one - qubit gates may be operated by nanoscale electromagnetic fields . where two - qubit gates are needed , neighboring q1dcs are allowed into a tunnel contact controlled by a potential on a surface gate.@xcite the use of the coulomb coupling between neighboring q1dcs is a common tool in spintronics which can be used to generate entangled states in dual - rail qubit representations.@xcite figure [ sawnetwork ] illustrates the network of q1dcs and qubit gates envisaged for performing a particular quantum computation . , width=325 ] the saw - trapped electron is well confined in all three spatial directions . in this paper we define cartesian coordinates such that the saw propagates along the @xmath8 axis with the @xmath9 axis normal to the 2deg . the 2deg is produced by a band - energy mismatch at the gaas / algaas interface which gives rise to a confining potential in the @xmath9 direction . the energy level spacings in the 2deg well are on the order of @xmath10.@xcite further confinement in the @xmath11 direction is provided by an extended q1dc , etched into the surface so as to avoid screening the saw - induced potential with metallic surface gates.@xcite finally , the confinement in the @xmath8 direction is due to the saw potential minimum which is approximately sinusoidal . the saw amplitude is typically 40 mev,@xcite which is sufficiently large to prevent qubits being lost via tunneling into neighboring saw minima . there are two important aspects of the saw quantum computation scheme that distinguishes it from other similar quantum dot schemes . first , the scheme provides repetitions of the same quantum computation with each passing of a single wavefront of the saw . therefore , a statistical time - ensemble of identical computations can be read out at the end of the q1dcs as a measurable current , alleviating the need for single - electron measurements . two sources of noise in the measured current may be estimated as follows : the shot noise is largely determined by how well the current is quantized to @xmath12 and precisions of @xmath13 can be experimentally achieved.@xcite johnson noise arises from the resistance of the ohmic contacts and the 2deg which are on the order of @xmath14 . at temperatures of @xmath15 they produce a rms voltage noise spectral density on the order of @xmath16 at most . this can not drive a current noise through the saw device since its effective intrinsic impedance of @xmath17 is comparatively very large . see experimental papers@xcite for more detailed discussions . the second key aspect of the scheme is the static nature of the gate components of the quantum circuit . this alleviates the need for strong , targeted and carefully - timed electromagnetic pulses that can be difficult and expensive to implement . the requirement of such expensive control resources often limit the scalability of most quantum computing implementations . it would certainly be convenient , though not essential , to have a means of preparing a pure fiducial qubit state . in nmr quantum computing , operations are carried out on ensembles of replica qubits which remain close to a highly mixed state of thermal equilibrium . nevertheless , a successful readout of the computation is obtained because the sum over many identical computations provides a measurable signal . similarly in the saw quantum computing scheme , states close to the maximally - mixed state are still useful because of the time - ensemble nature of the scheme . however , in contrast to nmr schemes , it is in principle possible to read out single electrons in the saw scheme . we will therefore begin with nonensemble quantum computation in mind and only later exploit the advantages of ensemble computation to deal with noise . of course , the more pure the qubit states remain , the faster the computation will converge to the result . for these reasons , we describe in this paper two simple methods for preparing pure fiducial qubit states . in the original proposal for the saw quantum computer,@xcite it was noted that the application of an external magnetic field of about @xmath18 will influence which spin states of electrons are favored in the capture process.@xcite this polarized capture process can be summarized as follows . the saw is strongly screened in the bulk 2deg until it is raised above the fermi energy and the quantum dot begins to form . when the higher - energy minority quantum dot forms for the higher - energy sub - band of polarized electrons , the probability of capturing minority electrons is small ( see fig . [ bcapture ] ) , whilst the probability of capturing electrons from the lower - energy sub - band is large . once a cloud of approximately polarized electrons is captured , the exchange interaction will generally entangle electrons in the same dot together , so that the subsequent loss of electrons into the fermi sea will lead to a decoherent process that could relax the remaining electrons into the low - energy polarized state . a more detailed multiparticle analysis will be needed to determine the final state of the remaining single electron , but the combination of the above two processes is likely to lead to a high level of polarization . is the fermi energy . the saw is strongly screened in the bulk 2deg below the fermi energy . as the saw enters the q1dc constriction , the confining potential begins to form . the probability of capturing spin - down is small at the point the minority spin - type dot forms ( upper curve ) . in contrast , the probability of capturing spin - up electrons is high when the majority spin - type dot forms ( lower curve ) . [ bcapture ] , width=325 ] the above method is conceptually simple and would be easy to implement in the laboratory . however , the macroscopic magnetic field required in the capture region will need to be shielded from the rest of the device where the quantum computation is to be carried out , and this may present a nontrivial problem . if we chose to drive spin flips using microwave pulses , then the macroscopic field is actually required across the whole device . however , the problem with using microwaves is their relatively long wavelength which tends to affect every part the computation . alternatively , we may use local , static magnetic fields to initialize , rotate and read out single qubits without a global magnetic field . in section [ section : readout ] , it will be demonstrated that spin - polarized electrons can be prepared and measured using a gate driven by the stern - gerlach effect . first we turn to the implementation of single qubit rotations using local static magnetic gates . single - qubit unitary operations may be carried out , in principle , by allowing the trapped single electrons to pass through regions of uniform magnetic field . if there is no spin - orbit coupling effect , the spin state of the qubit will evolve according to the zeeman term in the hamiltonian : @xmath19 , where @xmath20 is the magnetic field , @xmath21 is the vector form of the 3 pauli operators and @xmath22 is a spinor . the bloch vector @xmath23,2\ , \mathrm{re}[\alpha\beta^*],\|\alpha\|^2-\|\beta\|^2\right)$ ] precesses about the direction of the magnetic field with angular frequency @xmath24 . a local static magnet may be produced by a magnetic force microscope or by evaporative deposition of a ferromagnetic material such as cobalt , or a permalloy such as nife . inevitably , different samples will produce different strengths of magnetic field . but there are methods to vary the strength and pattern of the field once the sample has been fabricated . it has been demonstrated that ferromagnetic properties of thin - film 3d transition metals can be modified via ion irradiation.@xcite another method would be to use oxidation techniques with the atomic force microscope . only two independent directions of the b - field are necessary to produce an arbitrary single - qubit manipulation and we will choose these to be perpendicular to the direction of the saw , one aligned with and the other perpendicular to the 2deg . for idealized qubits with no spatial degree of freedom , the above model for single - qubit rotations is complete . however , the trapped electron is a charged particle with a spatial distribution within the dot . the fields couple to both the spatial and spin degrees of freedom , causing the electron to experience the lorentz force as well as spin - precession . it is therefore clear that one can not increase magnetic fields arbitrarily , since the lorentz force will upset the confinement properties of the electron . nor can the direction of the field be chosen arbitrarily without consequences for the robustness of the gate . to address these concerns , we analyze the behavior of the full wavefunction : a two - component spinor field defined over space and time , @xmath25 , under the action of gates operated by static magnets . assuming a uniform magnetic field in the lab frame , we solve the pauli equation for the spin field from which the probability density field and the bloch vector field can be obtained . the qubit is trapped in a net electrostatic potential with contributions from the q1dc split gates , the 2deg confining potential and the saw piezoelectric potential . the parameter regimes we consider allow us to neglect motion out of the 2deg plane . dependence of the wavefunction . this is partly justified if @xmath26 , where @xmath27 is the characteristic energy gap between the 2deg energy bands and @xmath28 is the length of the gate . this is indeed the case for parameter regimes that we consider . ] the net confining potential in the @xmath29 plane is modeled by @xcite @xmath30\big\}. \label{v}\end{aligned}\ ] ] the q1dc split gate voltages are such that typically @xmath31 . the width of the q1dc @xmath32 is typically between @xmath33 and @xmath34 . the amplitude of the saw is @xmath35 with wavelength @xmath36 and frequency @xmath37 . the appropriate nonrelativistic equation for the two - component spinor field is the schrdinger equation with a pauli hamiltonian . for an electron moving in an arbitrary vector potential @xmath38 and potential energy @xmath39 , the pauli hamiltonian @xmath40 is @xmath41 where @xmath42 is the electronic charge , @xmath43 is the land g - factor and @xmath44 is the effective mass in gaas . by simulating in the plane of the 2deg we simplify to a 2 + 1 dimensional model , with variables @xmath45 . if the particle were chargeless but with an anomalous magnetic moment , terms in @xmath40 that explicitly involve @xmath1 can be dropped . in such a case , there is no lorentz force acting on the particle , the schrdinger equation simplifies significantly and there would be no need to concern ourselves with the spatial behavior of the qubit . before proceeding further , it is convenient to transform into the rest frame of the electron moving with the saw speed @xmath46 . we may use simple galilean transformations @xmath47 , since the saw velocity is non - relativistic . the potentials in the electron rest frame become , in the harmonic oscillator approximation , @xmath48\\ & \simeq & \frac{v_0}{2 w^2}\ , { y'}^2 + \frac{a}{2}\ , k^2 { x'}^2 , \label{v'}\end{aligned}\ ] ] with @xmath49 . a further convenience is to use a system of natural units such that @xmath50 are unity . the units of length , time and energy become @xmath51 , @xmath52 and @xmath53 , respectively . the natural unit of magnetic field is @xmath54 . parameters can now be assigned with dimensionless values with respect to the above units . in the electron rest frame , @xmath55 \psi + { \frac{1}{2}}a^2 \psi \\ + { \frac{1}{2}}\ , g { \mu_\mathrm{b}}\ , \sigma\cdot b \psi + v \psi = i { \frac{\partial } { \partial t}}\psi , \end{split}\ ] ] where the primes will subsequently be dropped from the coordinates . we will further simplify the hamiltonian by adopting the coulomb gauge @xmath56 . for a magnetic field in the @xmath11 direction in the plane of the 2deg and transverse to the q1dc , we model the magnetic field in the region of the single - qubit gate with a vector potential of the form @xmath57 with @xmath58 . clearly this does not vanish at infinity , but we only consider interactions over regions of finite extent . this potential generates a uniform magnetic field of strength @xmath59 in the @xmath11 direction . with this potential eq . ( [ paulieq2 ] ) becomes @xmath60 in which the potential @xmath39 is given by ( [ v ] ) . an effective potential @xmath61 can be identified , which resembles a harmonic oscillator but is time - dependent . anticipating an analytic solution by separation of variables we apply the ansatz @xmath62 where @xmath63 are the harmonic oscillator eigenstates with energies @xmath64 , and @xmath65 are the eigenstates of @xmath66 with eigenvalues @xmath67 . we have introduced the oscillator frequency @xmath68 and the zeeman energy gap @xmath69 . the energy eigenstates in the @xmath11 direction are exactly the harmonic oscillator modes . the wavefunction in the @xmath8 direction is time - dependent and it is the solution of @xmath70 to which we now turn . on substituting ( [ ansatz ] ) into ( [ paulieq3 ] ) , we obtain a pde for @xmath70 : @xmath71\chi = i\ , \partial_t\ , \chi,\ ] ] with @xmath72 and @xmath73 . terms which depend only on @xmath74 have been dropped , as they merely contribute global , time - dependent phases that do not affect the dynamics . a gaussian solution of ( [ chiequation ] ) can be found with a further ansatz @xmath75.\ ] ] this system of time - dependent functions can be determined self - consistently , assuming a gaussian groundstate of the saw dot for the initial condition at @xmath76 . the resulting expressions for @xmath77 are complicated , but by noting a few general properties of the solution we can understand all the important features of the dynamics . @xmath78 takes account of the normalization but is otherwise of no more interest . @xmath79 and @xmath80 together describe a groundstate gaussian wavefunction evolving in the time - dependent vector potential . the solution is particularly simple in that it remains gaussian throughout , so we will only need to keep track of the position of the central peak @xmath81 and the width ( standard deviation ) @xmath82 of the probability distribution @xmath83 : @xmath84}{\mathrm{re}[f_1(t ) ] } \\ \delta(t ) & = & { \frac{1}{2}}\ , \sqrt{-\mathrm{re}[f_1(t)]}.\end{aligned}\ ] ] the typical energy scales for the saw electron encountered in experiment puts the system in the regime where @xmath85 and @xmath86 . expanding @xmath87 and @xmath88 to first order in @xmath89 and @xmath90 , we obtain the asymptotic behavior for the position of the peak : @xmath91 from this result we see that the magnetic field introduces a constant drift velocity of the peak in the @xmath92 direction . on exiting the interaction region , the peak will be off - center with respect to the saw dot and the probability distribution will subsequently oscillate in the @xmath8 direction , perhaps exciting higher - energy orbital states of the dot . if the charge distribution is pulled too far off - center , it is likely to escape the saw quantum dot . we could ask how long the electron can remain in the field before it is dragged a distance @xmath93 away from the center of the dot in the @xmath8 direction : @xmath94 when this is compared with the time required for the bloch vector to rotate by some appreciable angle such as a @xmath95 rotation , we obtain the ratio @xmath96 which is about @xmath97 for a @xmath18 field . this means that the bloch vector is allowed to precess approximately @xmath98 times before the drag effect destabilizes the qubit confinement . in the limit of weak fields , this ratio is no smaller than @xmath99 . therefore , whilst it is possible that higher orbital states are excited as a consequence of the interaction , there is little danger of the electron escaping the saw quantum dot during the operation of the gate . we now turn to the behavior of @xmath100 , the width of the gaussian wavefunction , which in the limit @xmath101 oscillates according to the simple expression @xmath102^{1/2},\ ] ] with @xmath103 and @xmath104 . the frequency of the oscillation increases with the energy of the dot and is typically much faster than the zeeman spin precession frequency , i.e. , rate of precession of the qubit bloch vector . these results are plotted for a specific case in fig . [ gaussiansolution ] . the top two plots show the evolution of @xmath81 and @xmath82 . as the bloch vector rotates about the field along the @xmath11-axis , the gaussian translates in the @xmath8 direction with a rapidly oscillating width . the bottom plot of fig . [ gaussiansolution ] shows probability amplitudes of excitation into higher - energy simple harmonic oscillator ( sho ) modes in the @xmath8 direction of the quantum dot . in a constant magnetic field @xmath105 . the top plot shows peak position @xmath81 and the middle plot shows the width @xmath82 . the bottom plot shows the probability amplitudes @xmath106 of the @xmath107th sho mode in the @xmath8 direction : @xmath108 ( long dash ) , @xmath109 ( short dash ) . the plot also shows @xmath110 ( solid ) . typical values for the saw dot are taken : @xmath111 ; @xmath112 ; magnetic field of @xmath18 . @xmath113 is the time taken for a @xmath95 rotation of the bloch vector about the @xmath11-axis . the initial gaussian wavefunction was taken to be the saw quantum dot groundstate @xmath114 . in the limit @xmath115 ( weak field ) , @xmath116 moves approximately linearly and @xmath100 undergoes bounded oscillations about @xmath117 . [ gaussiansolution ] , width=317 ] our conclusions are as follows . in addition to rotating the bloch vector as required , the gate has the effect of displacing the center of the gaussian wavefunction which will increase the energy of the bound state by an amount depending on both the gate time and the field strength . higher - energy orbital states are likely to be excited , which could lead to decoherence via spin - orbit couplings or dipole coupling to phonons and photons . according to fig . [ gaussiansolution ] , the electron is almost completely out of the groundstate after a @xmath118 rotation of the bloch vector . in the extreme case after many spin precessions ( typically thousands ) , the qubit will leave the saw quantum dot . however , for quantum computations the gate time need only be long enough to conduct a single orbit of the bloch sphere , in which case the excitation into higher orbital states is negligible . these observations reveal some important features of the gate operation that are not revealed by an idealized spin - only model of the qubit . to move the qubit state to an arbitrary point on the bloch sphere a second axis of rotation on the bloch sphere is needed , and to this end we consider a gate implemented by a uniform magnetic field in the @xmath9 direction . this will not be just a trivial extension of the preceding analysis , since the 3d rotation symmetry is broken by the motion of the saw in the @xmath8 direction . in practice , the @xmath11 magnetic field may be easier to fabricate than the @xmath9 magnetic field , since the latter passes perpendicularly through the 2deg structure . it could feasibly be produced by layering oppositely aligned thin - film magnets just beneath and just above the 2deg , or alternatively by applying a global @xmath9 magnetic field which is shielded in regions where it is not needed . a vector potential generating the uniform @xmath119 magnetic field is @xmath120 with @xmath121 . it should be noted that although the field is uniform and static in the laboratory frame , the electron sees a moving uniform field . in the electron rest frame , @xmath122 is time - independent and @xmath9-independent , allowing a hamiltonian in @xmath8 and @xmath11 only . the chosen gauge facilitates numerical simulations which follow shortly . an effective scalar potential that is quadratic in @xmath11 arises from the @xmath123 term in the hamiltonian ( [ paulieq2 ] ) . it is interesting to compare the strength of this confining potential with the q1dc potential , which is also approximately quadratic in y : @xmath124 with magnetic fields of order 1 tesla and typical q1dc energies , this ratio is of the order unity . this means that for the parameter values being considered the effective scalar potential arising form the @xmath123 term is comparable to the q1dc confinement potential . a stronger field would start to significantly deform the shape of the dot . this will not be a problem as long as the evolution has occurred adiabatically during the deformation . however , as we shall discuss in a moment , the probability of excitation into higher orbital states due to the @xmath119 field is not negligible . the other interesting term in the hamiltonian is the asymmetric coupling @xmath125 between the @xmath8 and @xmath11 variables . this is expected to introduce rotational behavior , as we would intuitively anticipate some form of landau orbital motion due to the lorentz force . a numerical simulation implementing a crank - nicholson , @xcite finite difference algorithm ( alternating direction method ) was used to simulate the operation of the gate . we started with an initial gaussian groundstate of the dot with spin state @xmath126 and subjected it to @xmath18 of magnetic field for a duration of @xmath127 , which is the gate time required for a @xmath118 rotation of the bloch vector about the @xmath9 axis . the evolution of the bloch vector is simple due to the uniformity of the magnetic field : @xmath128 , @xmath129 , and @xmath130 . all other parameters were assigned those values given just after eq . ( [ v ] ) . the result of the simulation is shown in fig . [ rotation : z ] , which shows time - shots of the probability distribution in the 2deg ( @xmath29 ) plane . losing its initial elliptic contours , the distribution develops two lobes which rotate about its midpoint . the density at the center increases due to the spatial squeezing from the @xmath123 term in the hamiltonian . this clearly shows excitation into higher - energy orbital states . using perturbation theory to second order in the strength of the field @xmath59 , with harmonic oscillator modes as the basis , we found the following facts : ( i ) to second order of perturbation theory , only the second excited state with sho quantum numbers @xmath131 becomes populated ; ( ii ) the probability ratio with respect to the groundstate is @xmath132}{\omega_y ( { \omega_x } + { \omega_y})^{2}},\ ] ] where @xmath133 and @xmath134 are frequencies arising from the harmonic oscillator approximation ; ( iii ) the amplitude of the ratio approaches unity when the field approaches @xmath18 ; ( iv ) the ratio of the spin precession frequency to the frequency of the @xmath135 oscillation is about @xmath136 for typical saw parameters ( see just after eq . ( [ v ] ) ) . the above model is useful in assessing the robustness of the qubit and its susceptibility to decoherence due to orbital motion . population into this excited state is a problem for decoherence , since the oscillating charge in the dot couples via dipole interactions to phonons and other charges outside the dot . however , provided we have sufficient control of the gate time , we can use ( [ oscpert ] ) to make the electron exit from the gate in its groundstate . otherwise , a gate driven by radiofrequency pulses in the presence of a global magnetic field could provide an alternative means to implement spin rotations about the @xmath9 axis . ( 2deg ) plane of a saw electron undergoing a @xmath118 rotation gate about the @xmath9 axis , under a uniform magnetic field @xmath119 of strength @xmath18 . snap - shots are shown at the following times ( ps ) : ( a ) 0 , ( b ) 2.89 , ( c ) 11.6 , ( d ) 20.3 , ( e ) 28.9 , ( f ) 37.6 . the initial probability distribution is a gaussian groundstate with standard widths @xmath137 and @xmath138 . the probability density rotates about the @xmath9 axis under the influence of the lorentz force acting in the @xmath29 plane . moreover , an effective scalar potential contributed by the @xmath123 term in the hamiltonian is quadratic in @xmath11 , which spatially squeezes the initial gaussian distribution into the center of the dot . [ rotation : z ] , width=316 ] in addition to single - qubit rotation gates , we require the ability to initialize and measure qubits at the beginning and end of the computation . spin - polarized electrons can be obtained from injection through a ferromagnetic contact.@xcite there is also a method to polarize spin using nondispersive phases ( aharonov - bohm and rashba ) without the need for ferromagnetic contacts.@xcite in the field of quantum computing , a well - known method for achieving readout of solid state spin qubits is to convert spin information into charge information@xcite and subsequently use single - electron transistors or point contacts to detect charge displacements or rabi oscillations . however , recent theoretical results by stace and barrett @xcite argue the absence of coherent oscillations in a continuously measured current noise , contrary to previous results and assumptions,@xcite and therefore raise concerns about the measurability of charge oscillations in similar scenarios . in any case , it is difficult to apply these methods to qubits in moving quantum dots . we therefore turn towards a quite different approach , one which enables the initialization and readout of saw electron spin qubits solely with the aid of nanomagnets and ohmic contacts.@xcite the readout gate we consider is based on the stern - gerlach effect . in the 1920s bohr and pauli asserted that a stern - gerlach measurement on free electrons was impossible,@xcite using arguments which combined the concept of classical trajectories and the uncertainty principle . this subsequently led physicists to analyze single - electron stern - gerlach measurements within increasingly more rigorous quantum settings , ultimately ending the debate by showing that the measurement can indeed be done , albeit with certain caveats . thus stern - gerlach measurements on _ free _ electrons have been extensively investigated , but little attention has been given to such measurements on confined electrons.@xcite an interesting semiclassical analysis of a stern - gerlach type experiment with conduction electrons has been reported,@xcite in which the authors justifiably neglect the lorentz force effects . in contrast , we analyze a single - electron stern - gerlach device , providing a full quantum mechanical treatment and including all lorentz force effects . in the saw electron system , single electrons are transported in moving quantum dots . can this system undergo a stern - gerlach type measurement , and would it be of any use in a scheme for quantum computation using qubits trapped in surface acoustic waves ? the electron confinement to low dimensions allows us to guide the electron through the magnetic field in ways that enhance the spin measurement and suppress the deleterious effects of delocalization and the lorentz force . surface magnets can be arranged in such a way as to produce a local magnetic field inhomogeneity . for example , two north poles placed on either side of the q1dc will produce a region of intense magnetic field gradient inbetween ( see fig . [ sgdevice ] ) . via the zeeman interaction term @xmath139 in the pauli hamiltonian , the field inhomogeneity has the effect of correlating the spatial location of a wavepacket to its spin state.@xcite to partially delocalize the particle during the gate operation ; @xmath140(@xmath141 ) labels the q1dc receiving electron flux in the spin up(down ) state of the @xmath66 operator ; magnets ( of any geometry and polarity ) need to produce a localized , inhomogeneous distribution of magnetic field . [ sgdevice ] , width=325 ] in most situations it is necessary to continue confining the qubit during the operation of the gate , because the spreading time for a free gaussian wavefunction is comparatively short on the order of @xmath142 . but the gate must cause a wavepacket splitting in order for the spin states to be resolved , hence the q1dc must relax to allow for motion in the @xmath11 direction . this is achieved by patterning the q1dc in a funnel shape , with an opening angle @xmath143 ( see fig . [ sgdevice ] ) , such that in the electron rest frame the potential looks like @xmath144 ^ 2,\ ] ] where @xmath145 is the step function : @xmath146 and @xmath147 . a negatively biased surface gate placed on the @xmath8 axis can be used to guide the electron into the channels . if necessary , both the position of the electrode and the opening angle @xmath143 could be optimized for a particular sample device under low - temperature and high - vacuum conditions by erasable electrostatic lithography.@xcite we will study the quality of the readout obtained from the gate for two different magnetic field configurations , both sufficiently simple so as to be realizable in the near future : the linearly inhomogeneous field and the 2d dipole field . in the first model we will analyze a simple , unidirectional and linearly inhomogeneous field pointing in the @xmath11 direction : @xmath148 with @xmath149 . a wedge - shaped single domain surface magnet of appropriate dimensions can produce an inhomogeneous magnetic field of this form near its center of symmetry . a vector potential for this field is @xmath150 , with @xmath121 . this field exerts a spin - dependent force in the @xmath11 direction and it is the simplest field that induces the stern - gerlach effect . although a @xmath9 dependence enters into the vector potential , by considering motion only in the @xmath151 plane we may avoid contributions from terms involving @xmath38 in the hamiltonian , as well as the @xmath9 component of magnetic field . the absence of @xmath8 in the potential immediately allows us to write down harmonic oscillator modes for the @xmath8 dependence . the remaining @xmath152 dependent part obeys @xmath153 where @xmath154 is the eigenvalue @xmath155 of the spinor @xmath65 , and @xmath156 is given by ( [ vq1dc ] ) . the initial state is again the gaussian ground state of the q1dc . a solution to ( [ diffeq ] ) is obtained by a time - dependent gaussian ansatz ( [ gaussansatz ] ) @xmath157.\ ] ] in a similar way as before , we solve the system of coupled ordinary differential equations and derive the time dependence of the standard deviation @xmath100 and the position @xmath116 of the probability distribution . let @xmath158 and @xmath159 . the width of the initial gaussian groundstate wavefunction is determined by @xmath160 . the behavior of the gaussian solution is then characterized by the following time - dependent parameters @xmath161 \\ \mu(\tau ) & = & \frac{\sqrt{\tau}\,c_1}{2\,(2\,\alpha^2+c_0 ) } \big [ \tau^{3/2 } - \cosh ( \sqrt{\gamma}\,\ln(\tau)/\alpha ) \nonumber\\ & & - \frac{3\alpha}{\sqrt{\gamma } } \sinh ( { \sqrt{\gamma}\,\ln(\tau)/\alpha } ) \big ] , \label{muevolution}\end{aligned}\ ] ] where we have introduced ancillary variables @xmath162 , @xmath163 and @xmath164 . the parameter @xmath165 is useful in determining when the trigonometric functions become oscillatory . if the angle @xmath143 is critical such that @xmath166 , then @xmath167 and the width has an especially simple behavior : @xmath168.\ ] ] this model predicts the width and position of the qubit wavefunction after a time @xmath74 in the linearly inhomogeneous field with field gradient @xmath59 . in practice we can not produce arbitrary magnetic fields in arbitrary configurations . for a weak magnetic field such that @xmath169 , @xmath170.\end{aligned}\ ] ] the width increases as @xmath171 and the position moves as @xmath172 as expected . we may compare @xmath100 with the dispersion of a gaussian in free space @xmath173 which is linear in @xmath74 at large times . the broadening of the width is suppressed due to the confinement potentials . the effect of the stern - gerlach can be pictured as follows . consider a qubit in the state @xmath174 entering the gate . the q1dc broadens as the qubit enters the field , allowing the stern - gerlach effect to produce a spin - dependent shift in the center of mass towards the channel @xmath140 . however , the initial localized distribution will deloclize due to the q1dc broadening , allowing a small current to be detected in the wrong channel @xmath141 . thus this gate is a spin polarizing filter with some intrinsic error rate , independent of decoherence effects , shot noise and johnson noise . it remains for us to judge how well it performs , and to this end we introduce a fidelity measure of the gate . as an input to the stern - gerlach gate , we consider a density matrix of the form @xmath175 where @xmath176 is the gaussian groundstate of the dot and @xmath177 is the initial spin density matrix . in the inhomogeneous field the spatial degrees of freedom couple to the spin degrees of freedom , leading to @xmath178 after a gate time @xmath74 , we post - select the state in , say , the @xmath140 channel and renormalize : @xmath179 the probability yield in the channel is the normalization @xmath180 of the post - selected state . the simplest projection kernel projecting into the support of @xmath140 is @xmath181 , where @xmath182 is the step - function defined in ( [ vq1dc ] ) . we finally define our fidelity as @xmath183 the yield @xmath180 can be calculated without explicit knowledge of @xmath184 , since the gaussian solutions are related by reflection symmetry : @xmath185 . we choose various magnetic field gradients ranging from @xmath186 to @xmath187 and plot the output fidelity of a maximally - mixed input state as a function of the q1dc angle @xmath143 and the gate time @xmath74 ( see fig . [ fidelitycontour ] ) . perfect spin filtering corresponds to a fidelity @xmath188 , and the worst - case fidelity is @xmath189 . the density plots show how the optimal parameter regions increase with larger magnetic field gradients . this model is useful in determining the appropriate gate geometry and field gradients required to achieve good fidelities . in the @xmath140 channel for a stern - gerlach gate operated by a linearly inhomogeneous and unidirectional magnetic field in the @xmath11 direction . white regions @xmath190 @xmath191 . @xmath143 is the angle between the q1dc and saw direction . @xmath74 is the gate time . input state is a maximally - mixed spin state in the groundstate of the saw quantum dot . various magnetic field gradients have been chosen to show the increase in the optimal parameter region ( white ) in the @xmath192 plane with increasing field gradient . @xmath193 is a crude order of magnitude estimate for the gaas spin relaxation lifetime . [ fidelitycontour ] , width=326 ] the technology of nanoscale single domain magnets is not yet capable of producing large fields in arbitrary configurations . the above simple linearly inhomogeneous magnetic field is therefore a simple approximation to the real magnetic field configuration on any particular device . in the next subsection we consider the effect of a different ( perhaps more realistic ) field configuration , namely , the field near one of the poles of a dipole magnet . in this subsection we study the quality of the readout obtained from a stern - gerlach type gate driven by a simple 2d dipole field lying in the 2deg plane , as shown in fig . [ vectorfield ] . in appendix [ derivedipolefield ] , we derive the magnetic vector potential @xmath194 where @xmath59 is a strength parameter and @xmath195 is the distance of the 2d dipole from the q1dc . ( 2deg ) plane showing the magnetic field due to an infinite string of dipole moments running parallel to the @xmath9-axis at a distance @xmath195 from the q1dc . [ vectorfield ] , width=345 ] figure [ vectorfield ] shows that the gradient of the field in the @xmath11 direction reverses direction at two points , @xmath196 . in order to prevent the moving qubit from experiencing opposite field gradients , we will restrict the gate time to @xmath197 so that the qubit moves in the region @xmath198 in the laboratory frame . the question of how we are to shield the extraneous field from the qubit does not have a simple solution . it has been suggested that a superconducting material such as niobium could be used to shield the magnetic field where it is not needed.@xcite the results of a 1d simulation of the wavefunction along the @xmath199 direction are shown in fig . [ dipolefieldsimulation ] . the numerical method used was a crank - nicholson algorithm adapted to the pauli equation . the plot shows three time - shots of the probability density @xmath200 , normalized to peak . the initial state is a gaussian groundstate with spin state @xmath174 . the simulation shows no significant spatial displacement of the electron density in the @xmath11 directions , in contrast to the behavior exhibited under the linearly inhomogeneous field ( c.f . ( [ muevolution ] ) ) . this is partially due to the effect of the lorentz force generated by the @xmath123 term in the hamiltonian , which contributes an effective potential that is highly confining in the @xmath11 direction . the @xmath123 confinement is so strong that the q1dc opening angle @xmath143 becomes irrelevant . bohr and pauli s claim that the lorentz force washes out the stern - gerlach effect are upheld in this particular scenario . another source of problems for this field configuration is that the @xmath201 states are not eigenstates of the field as they were with the linearly inhomogeneous field . figure [ blochvecosc ] shows the evolution of the bloch vector in a spin - only model of the electron spin qubit . it demonstrates how the bloch vector begins precessing about the @xmath8 component of the magnetic field , causing the stern - gerlach force to act in different directions . although careful timing of the gate and possible compensation mechanisms could be put in place , this gate is essentially not robust . we draw the conclusion that in parameter regimes relevant to single - electron transport by saw in gaas , a field of this type will not be suitable for a stern - gerlach measurement gate , and that field unidirectionality is generally an important requirement for quantum gates driven by static magnetic fields . is the direction transverse to the q1dc . this is a stern - gerlach type readout gate using the 2d dipole field ( @xmath202 at @xmath203 ) shown in fig . [ vectorfield ] . initial condition is a gaussian wavepacket with spin @xmath204 . sideways translation of the probability density is suppressed due to both the strongly confining effective potential arising from the @xmath123 term and the precessional motion of the bloch vector in the non - unidirectional field . [ dipolefieldsimulation ] , width=325 ] at @xmath203 ) : @xmath205 ( long - short dash ) , @xmath206 ( solid ) , @xmath207 ( short dash ) . time @xmath76 corresponds to the starting point @xmath208 ( see fig . [ vectorfield ] ) . we have used a spin - only model of the electron spin qubit to illustrate spin precession in the non - unidirectional field . the bloch vector changes direction rapidly , which in turn causes the stern - gerlach force to fluctuate during the operation of the gate . [ blochvecosc ] , width=302 ] the major effects of decoherence on this saw quantum computing system have already been considered qualitatively in the original proposal and subsequent papers,@xcite and more specifically by a number of authors . these effects include : the interaction of the electron qubit with other electrons in the 2deg , surface gates and donor impurities;@xcite the coupling of qubits to phonons,@xcite nuclear spins@xcite and radio - frequency photons@xcite . the models we have presented here give rise to behavior which was previously neglected in the original proposal for the quantum computation scheme.@xcite they allow us to predict the contribution to decoherence from excitation into higher orbital states and give their probabilities of occupation . any potential decoherence due to tunneling between neighboring dots is negligible , and if needed the spatial separation of qubits can be increased by either introducing higher harmonics of the saw fundamental frequency or by active gating at the entrance to each q1dc . a major source of error is the current fluctuations in the quantized single - electron saw current . in recent experiments,@xcite error in the current quantization ( including shot and johnson noise ) of less than @xmath209 were observed . the relevant decoherence timescales that underpin our proposals is the spin lifetime ( @xmath210 and @xmath211 ) of single electrons in quantum dots . for n - type bulk semiconductors , @xmath210 spin lifetimes in gaas of @xmath212 have been reported , which gives a very crude ball - park figure.@xcite an estimate that is more relevant to the saw electron system is the spin relaxation lifetime of a few nanoseconds , which was obtained from spin - resolved microphotoluminescence spectroscopy measurements on photoexcited electrons.@xcite however , this is the worst - case scenario because the method of electron capture we propose@xcite produces a conduction - band hole that is more extended , short - lived and mixed than the photoexcited hole . indeed , recent experiments on static gaas quantum dots report a lower bound on @xmath210 as long as @xmath213,@xcite which makes an estimate of @xmath212 for the @xmath210 lifetime of saw electrons quite reasonable in the light of current technology . we have analyzed in detail a proposal@xcite for implementing quantum computation on electron spin qubits trapped in saw - q1dc electrostatically - defined dots , using only static magnetic gates to perform single - qubit initializations , rotations and readouts . applying the full pauli hamiltonian , we described the quantum dynamics of both the spin and orbital states of the qubit for various parameter regimes relevant in saw single - electron transport experiments . in the analysis of single - qubit unitary operations with localized uniform magnetic fields , we showed that the effect of the lorentz force puts upper bounds on field strengths and gate times . moreover , simulations showed that field directions normal to the 2deg excite rotational states in the dot . probabilities of excitation into higher - energy orbitals were given . in terms of feasibility , the models indicate that a field strength of about @xmath214 will be sufficient to conduct a @xmath95 rotation in about @xmath215 , without compromising the confinement properties of the trapped qubit . since @xmath210 spin lifetimes of microseconds have been reported,@xcite it is feasible for hundreds of single - qubits gates to operate before all coherence is lost in the computation . we also studied a device for single - qubit measurement and initialization based on the stern - gerlach effect . the problematic lorentz force can be partly suppressed by virtue of the geometry of the 2deg system , and for a unidirectional , linearly inhomogeneous field , the correlation between the spin states and spatial location of the qubit leads to a good quantum measurement of its spin . for a 2d dipole field , the vector potential has a deleterious effect . namely , it contributes an effective confining potential via the @xmath123 term in the pauli hamiltonian which suppresses the transverse motion in the @xmath11 direction . furthermore , the component of magnetic field in the @xmath8 direction causes the bloch vector to rotate in undesired directions . the latter problem could be overcome by using unitary gates prior to the stern - gerlach gate to correct any spurious rotations , but it is unlikely that such a finely - tuned arrangement can be made robust . magnetic fields that are good approximations to a unidirectional and linearly inhomogeneous field can provide a source of polarized electron qubit states in channels @xmath140 and @xmath141 with high yield . the very same gate can be used to measure the ratio @xmath216 for an input qubit state @xmath217 . the advantages of this readout method are that the averaging time of the measured current can be made long enough to render most noise sources ( e.g. shot noise , input - referred noise of the current pre - amp ) unimportant , and that missing electrons from saw minima do not contribute any error since we are measuring the ratio of currents out of the two channels . field gradients in the range @xmath218 lead to gate times that are easily within the range of @xmath210 spin lifetimes in gaas . these parameter regimes will need to be probed by experiment if we are to make progress towards feasible quantum computation with nanomagnets in the saw quantum computer . the authors would like to thank sean barrett , masaya kataoka , alexander moroz , andy robinson and tom stace for their ideas and helpful discussions . the authors gratefully acknowledge the support of the epsrc . this work was partly funded by the cambridge - mit institute . we adopt cylindrical coordinates @xmath219 about the @xmath11 axis , with @xmath220 . we consider the field due to an infinite string of dipoles passing parallel to the @xmath9-axis through @xmath221 as shown in fig . [ vectorfield ] . the resulting magnetic field has planar symmetry with respect to the @xmath29 ( 2deg ) plane . the q1dc runs along the @xmath8 axis . a vector potential for this field can be constructed as follows . a magnetic dipole at @xmath222 pointing in the @xmath223 direction has a vector potential@xcite @xmath224 where @xmath225 and @xmath226 is the dipole moment per unit length in the @xmath9 direction . we now displace this potential by @xmath195 in the @xmath11 direction and integrate over all @xmath9 , obtaining @xmath227 with @xmath58 and strength parameter @xmath228 . the magnetic field can be derived by applying the curl in cylindrical coordinates : @xmath229 and @xmath230 . figure [ vectorfield ] shows the vector field plot of the resulting magnetic field , which falls off as @xmath231 for large @xmath232 .	in the surface acoustic wave quantum computer , the spin state of an electron trapped in a moving quantum dot comprises the physical qubit of the scheme . via detailed analytic and numerical modeling of the qubit dynamics , we discuss the effect of excitations into higher - energy orbital states of the quantum dot that occur when the qubits pass through magnetic fields . we describe how single - qubit quantum operations , such as single - qubit rotations and single - qubit measurements , can be performed using only localized static magnetic fields . the models provide useful parameter regimes to be explored experimentally when the requirements on semiconductor gate fabrication and the nanomagnetics technology are met in the future .
quantum information science ( qis)@xcite is a rapidly emerging discipline with the potential to revolutionize measurement , computation and communication . sitting at the intersection of quantum electronics , quantum optics , and information theory , qis offers a new paradigm for the collection , transmission , reception , storage and processing of information , based on the laws of quantum , rather than classical , physics . applications of qis certainly have the potential to generate totally new quantum information technology ( qit ) , but ultimately any future qit industry will be justified by commercial interest in the products and services that it supports . practical results in qis are currently right at the cutting edge of experimental quantum research , so the route to qit is very hard going . from the commercial perspective , a major challenge is the creation of relatively simple qit which is nevertheless economically viable . this would generate revenue and expand current industrial participation and interest in the field , effectively seeding a new qit industry , in the same way hearing aids were the first commercial application of the transistor and the beginning of the classical it industry . a long term goal for qit is the realization of many - qubit scalable quantum processors . it is known that these machines would outperform their classical counterparts at certain tasks such as factoring@xcite and searching,@xcite and the search continues for new applications . a shorter term goal is the realization of ( say ) 50 - 100 qubit processors . these would certainly be better at quantum simulation@xcite than any foreseeable conventional it and , as a research tool , would expose qit to a whole new class of curious and creative people , with all their potential for new ideas and applications . however , perhaps the most immediate qit , that which will stimulate a new industry , is based on or related to quantum communication@xcite and metrology.@xcite it has become clear over the last two decades that computer and network security that relies primarily on software protocols is potentially porous , being based on unproven mathematical assumptions . in qis , quantum computation and communication protocols can be devised in which unconditional privacy is guaranteed by fundamental laws of physics . although it is not clear yet that quantum key distribution ( qkd ) will be the first profitable application of qit , it is possible that extensions of qkd ( e.g. , controlled entanglement swapping ) , photonic state comparison ( for quantum signature verification ) , and full quantum communication at high data rates will become compelling to financial , medical , and other institutions and their customers . furthermore , it is already clear that distributed quantum algorithms can efficiently enable solutions to economics problems ( e.g. , public goods economics@xcite ) that are difficult to treat with conventional mechanisms , but it is not yet known whether other economic procedures such as auctions have superior quantum solutions . similarly , quantum metrology and imaging have interest for the nanoscale manufacturing and physical security industries , as these techniques allow tiny phase shifts , displacements , and forces to be accurately measured remotely even when the target is enclosed within an inaccessible or hostile environment . although there are still open questions , one promising route for starting a qit industry is to found it on communication and metrology applications , based on the generation , transmission , processing , and detection of a few photonic qubits . it is certainly clear that photons ( or other quantum states of light ) are _ the _ qubits of choice for communication,@xcite and so for few - qubit processing there is a case for keeping everything optical . it is also the case that potentially useful processing tasks could be performed with moderate ( 1020% ) gate error rates , rather larger than the stricter error bounds demanded for fault - tolerant many - qubit processing . this `` all optical '' scenario is the motivation for our work . clearly , taking this approach , quantum information processing primitives based on nonlinear quantum optics ( such as a universal set of optical gates and single - photon detectors ) must be developed and fabricated . these primitives would potentially allow the construction of few - qubit nanoscale quantum optical processors that could be incorporated into existing pcs and communication networks . in this paper we discuss the possibility of realizing these primitives through use of electromagnetically induced transparency ( eit ) . atom and a nearly resonant three - frequency electromagnetic field . ( a ) in the semiclassical view , two atomic energy levels are separated by the energy @xmath1 , and coupled by a field oscillating at the frequency @xmath2 . the strength and phase of the corresponding dipole interaction is represented by the rabi frequency @xmath3 . ( b ) in the quantum view , the states of the atom + photons system separate into manifolds coupled internally by resonant transitions . ( c ) a model mach - zehnder interferometer illustrating an architecture for a `` dual rail '' quantum phase - shifter using four - level @xmath4 atoms . the upper arm is denoted by `` 1 '' and the lower arm by `` 0 . '' ] in previous work,@xcite we considered a model of the nonlinear electric dipole interaction between three quantum electromagnetic radiation fields with angular frequencies @xmath5 , @xmath6 , @xmath7 and a corresponding four - level @xmath4 atomic system , as shown in fig . [ fourlevel](a ) . we considered @xmath8 atoms , fixed and stationary in a volume that is small compared to the optical wavelengths , and we assumed that the three frequency channels of the resonant four - level manifold of the resulting quantum system shown in fig . [ fourlevel](b ) are driven by fock states containing @xmath9 , @xmath10 , and @xmath11 photons , corresponding to the rabi frequencies @xmath12 , @xmath13 , and @xmath14 , respectively . as an example of the use of an eit system as a phase - shifter , we incorporate the atomic system into the dual - rail mach - zehnder interferometer shown in fig . [ fourlevel](c ) . we wish to apply a phase shift to the photon in mode @xmath15 on the upper rail , conditioned on the presence of one or more photons in mode @xmath16 . in one arm of the interferometer , the four - level atoms are prepared using @xmath17 to provide a phase shift at the probe frequency @xmath5 while remaining largely transparent and dispersive . in the second arm , @xmath18 , and the system is tuned to match the absorption and dispersion provided by the atoms in the first arm , allowing the interferometer to remain time - synchronous . we must be careful to demonstrate that the interaction of either arm with a photon at the probe frequency that has entered the interferometer at the input port will entangle the optical modes with each other but _ not _ with either collection of atoms . therefore , we solve the density matrix equation of motion in the presence of a completely general lindblad damping model , and monitor the element @xmath19 that corresponds to the initial ( ground ) state of the entire collection of atoms in both ensembles . if the field in mode @xmath20 is indeed described by a fock state , then using the quasi - steady - state approximation to obtain eqs . [ eq : rho10 ] and [ eq : w_10_4 ] . however , our subsequent numerical work has shown us that those results hold over a broad range of experimental parameters even when @xmath21 . ] ( @xmath22 ) we obtain@xcite @xmath23 where the rabi frequencies @xmath24 and detunings @xmath25 are defined in fig . [ fourlevel](a ) and ( b ) , and @xmath26\ , \left\|\omega_a\right\|^2 } { ( \nu_a + i \gamma_{20})\left[(\nu_a - \nu_b + i \gamma_{30})(\nu_a - \nu_b + \nu_c + i \gamma_{40 } ) - \left\|\omega_c\right\|^2\right ] - ( \nu_a - \nu_b + \nu_c + i \gamma_{40 } ) \left\|\omega_b\right\|^2 } .\ ] ] the constant @xmath27 represents the net decoherence rate ( depopulation + dephasing ) of level @xmath28 of the quantum manifold shown in fig . [ fourlevel](b ) for the atom + field system on the first rail relative to the evolution of a system ( absent mode @xmath16 ) on the second rail , while @xmath29 represents the decoherence rate of the collective ground state divided by the number of atoms @xmath8 . since the atomic levels @xmath30 and @xmath31 are metastable by assumption , the decoherence rates @xmath29 and @xmath32 represent pure dephasing mechanisms . suppose that we consider the concrete case of a phase gate that couples single photons in modes @xmath15 and @xmath16 and is driven by a coherent state in mode @xmath20 . we wish to optimize the experimentally controllable parameters so that we introduce a phase shift between the two arms of the system with minimum error . we proceed by choosing _ a priori _ an error @xmath33 which occurs over the entire gate operation , and determining the parameters needed for the gate to perform with this level of error . there will be three sources of error : the dephasing described by @xmath29 and @xmath32 ; the additional depopulation described by @xmath34 and @xmath35 ; and the error arising from the finite value of @xmath36 for the coherent state in mode @xmath20 , which prevents the system from evolving to a perfect phase - shifted state even in the absence of decoherence . suppose that a phase shift of @xmath37 is applied by the phase gate after a time @xmath38 . then the density matrix element given by eq . ( [ eq : rho10 ] ) can be rewritten as @xmath39 , where the effective decoherence time @xmath40 is defined as @xmath41 for a given gate fidelity @xmath42 ( a measure of the distance between the ideal output state of the gate and the state that is actually generated@xcite ) , the net error @xmath43 due to dephasing and depopulation is @xmath44 . in the absence of significant dephasing , it is clear that a large value of the detuning @xmath45 reduces the value of the ratio @xmath46 and therefore @xmath40 ; in fact , the maximum practical value of @xmath45 would be set by the duration of the pulses in the three electromagnetic modes and the experimental convenience of accurately measuring the detuning . however , when the pure dephasing terms @xmath29 and @xmath32 are finite , @xmath40 reaches a minimum for a finite value of @xmath45 . we assume that the dephasing rates are small enough that @xmath47 ) , we obtain @xmath48^\frac{1}{2}\ , \frac{q_c^2}{q_b^2 } \qquad \mathrm{and } \qquad \tau_\mathit{eff } = 2 \left[\tilde{\gamma}_{10}\tilde{\gamma}_{20}\left(\gamma_{10 } \tilde{\gamma}_{20 } + \left\|\omega_a\right\|^2\right)\right]^{\frac{1}{2 } } \frac{q_b^2}{\left\|\omega_b\right\|^2 } \frac{q_c^2}{\left\|\omega_c\right\|^2 } \frac{\phi}{\left\|\omega_a\right\|^2 } , \ ] ] where @xmath49 consider the case of a @xmath0 phase gate , and assume that the pure dephasing rates are small enough that @xmath50 . we see immediately that we can have @xmath51 only if @xmath52 , giving @xmath53 and @xmath54 . now , if @xmath55 , then we must have @xmath56,@xcite this requirement is eased considerably to @xmath57 is replaced by the coherent state @xmath58 with @xmath59 ) shows that @xmath60 , and it is clear that the gate performs a phase shift when there is a photon present in _ both _ modes @xmath15 and @xmath16 , and no phase shift when either is absent . however , mode @xmath20 is driven by a coherent state , and an input coherent state will evolve according to@xcite @xmath61 where @xmath62 , @xmath63 make a significant contribution to the sum , and the phase can be approximately factored out as @xmath64 . however , if the coherent state is too weak , then the phase shift will be imperfectly implemented with a significant error , even in the absence of decoherence . furthermore , we see from eqs . ( [ eq : gt ] ) and ( [ eq : g20_t ] ) that the effect of the decoherence increases with the intensity of the coherent state , and therefore for a given dephasing there is some ideal value of @xmath36 that minimizes the error.@xcite we are now ready to estimate the magnitude of the dephasing that must be achieved to obtain a given error . we assume that @xmath21 and @xmath65 . we obtain @xmath66 this equation gives an estimate of the error arising from a _ fock _ input state with an optimized value of @xmath45 for a particular level of decoherence . in fig . [ fig : gplot](a ) we plot the error induced on a _ coherent _ input state due to both decoherence and the finite intensity of the coherent state as a function of the dephasing @xmath29 , where we have chosen both @xmath45 and @xmath67 to minimize the error . figures [ fig : gplot](b ) and [ fig : gplot](c ) shows these values of @xmath36 and @xmath45 chosen to give the largest possible dephasing for a given error . is not suppressed , i.e. @xmath68 . the dashed lines ( - - - ) illustrates the case when @xmath35 has been suppressed relative to @xmath34 by a factor of 1000 . the dotted lines ( @xmath69 ) give the error in the case of the one - qubit phase gate when modes @xmath20 and @xmath16 are occupied by coherent states . in ( c ) , the solid and dashed lines correspond to an ideal value of @xmath67 . however , the dotted line ( @xmath69 ) , shows a _ minimum _ value that @xmath67 must take for the error in the one - qubit gate to remain below the value shown ; @xmath67 can take on larger values without affecting either the decoherence or the value of @xmath45 required to minimize the error . for the one qubit gate , @xmath70 . ] as noted above , if @xmath35 can be suppressed , then the appropriate constraint @xmath71 becomes relatively easy to satisfy . for the requirement @xmath72 , the dephasing can be larger than in the unsuppressed case by a factor of @xmath73 and still produce the same error . by contrast , for @xmath74 , the dephasing can be larger by a factor of @xmath75 . therefore , suppressing @xmath35 means that the detuning required is also much smaller . in order to perform a two - qubit conditional phase gate with an error of 20% ( and assuming we can not suppress the depopulation from level @xmath76 ) , from fig . [ fig : gplot ] we find that the required parameters are @xmath77 , @xmath78 , and @xmath79 , and the gate operation time is determined by @xmath80 . however , if @xmath35 can be suppressed by a factor of 1000 relative to @xmath34 , the requirements are much reduced : the dephasing needed is @xmath81 , the detuning is @xmath82 , and the coherent state must have @xmath83 . the corresponding gate operation time is determined by @xmath84 . this system can also be used as a one - qubit phase shift gate ( or an eit - based qnd detector,@xcite where we can assume that @xmath85 ) , _ without _ the extra effort described above for suppression of spontaneous emission from the atomic level @xmath86 . if mode @xmath16 remains in the single - photon fock state , the system will act as a phase shifter on mode @xmath15 , and the above analysis applies . alternatively , we can replace the quantum state in mode @xmath16 with an intense coherent - state driving field . now moderate ( i.e. , non - classical ) values of @xmath67 and @xmath87 introduce errors . however , examining eqs . ( [ eq : g20_t ] ) , we see that unlike for the two - qubit gate the effects of decoherence depend much less sensitively on the intensities of the coherent states . as shown in fig . [ fig : gplot ] , this means that we can eliminate the error due to the finite size of the coherent state , and we are only left with an error due to decoherence . we have studied gates for quantum information processing based on a quantum treatment of eit systems . we have analyzed in detail the performance of a two - qubit phase gate ( and , by extension , that of a one - qubit phase gate ) as functions of both atom and field properties , and we have described a general optimization method that selects a detuning that minimizes the gate error for a given phase shift . the resulting constraints on the allowable dephasing rates in these systems are quite stringent for high - fidelity operation . however , if the spontaneous emission rate from the atomic level @xmath88 can be suppressed significantly , then demonstration of a moderate - fidelity phase gate becomes experimentally achievable . 99 m. a. nielsen and i. l. chuang , * quantum computation and quantum information * ( cambridge university press , 2000 ) . p. w. shor , `` polynomial - time algorithm for prime factorization and discrete logarithms on a quantum computer , '' proc . 35th annual symposium on the foundations of computer science , ed . s. goldwasser , 124 ( ieee computer society press , los alamitos , ca , 1994 ) ; siam j. computing * 26 * , 1484 ( 1997 ) ; quant - ph/9508027 . l. k. grover , `` a fast quantum mechanical algorithm for database search , '' proc . 28th annual acm symposium on the theory of computing ( stoc ) , 212 ( may 1996 ) ; quant - ph/9605043 ; phys . rev . lett . * 79 * , 325 ( 1997 ) ; quant - ph/9706033 . s. lloyd , `` universal quantum simulators , '' science * 273 * , 1073 ( 1996 ) . n. gisin et al . , `` quantum cryptography , '' rev . phys . * 74 * , 145 ( 2002 ) . h. lee , p. kok and j. p. dowling , `` quantum imaging and metrology , '' proceedings of the sixth international conference on quantum communication , measurement and computing ( qcmc02 ) , edited by j.h . shapiro and o. hirota ( rinton press , paramus , new jersey , usa , 2003 ) , quant - ph/0306113 .. k .- y . chen , t. hogg and r. g. beausoleil , `` a quantum treatment of public goods economics , '' quant . * 1 * , 449 ( 2003 ) ; arxiv : quant - ph/0301013 . r. g. beausoleil , w. j. munro , and t. p.spiller ( 2003 ) , `` applications of coherent population transfer to quantum information processing , '' j. mod . opt * 51 * , 1159 ( 2004 ) ; arxiv : quant - ph/0302109 . w. j. munro , k. nemoto , r. g. beausoleil , and t. p.spiller , `` a high - efficiency quantum non - demolition single photon number resolving detector , '' arxiv : quant - ph/0310066 .	we provide a broad outline of the requirements that should be met by components produced for a quantum information technology ( qit ) industry , and we identify electromagnetically induced transparency ( eit ) as potentially key enabling science toward the goal of providing widely available few - qubit quantum information processing within the next decade . as a concrete example , we build on earlier work and discuss the implementation of a two - photon controlled phase gate ( and , briefly , a one - photon phase gate ) using the approximate kerr nonlinearity provided by eit . in this paper , we rigorously analyze the dependence of the performance of these gates on atomic dephasing and field detuning and intensity , and we calculate the optimum parameters needed to apply a @xmath0 phase shift in a gate of a given fidelity . although high - fidelity gate operation will be difficult to achieve with realistic system dephasing rates , the moderate fidelities that we believe will be needed for few - qubit qit seem much more obtainable .
soft gamma - ray repeaters ( sgrs ) and anomalous x - ray pulsars ( axps ) have been known as strongly magnetized neutron stars , _ magnetars_. the hallmark of these objects is to repeat x - ray or gamma - ray emissions irregularly . so far , four or five known active objects which show frequent x - ray emission of tremendously energetic and shorter initial bursts ( typically @xmath0 ergs and @xmath1 s ) have been identified with sgrs ( thompson _ et al . some of them occasionally exhibit more energetic events . the giant flare 1900 + 14 , now associated with sgr 0525 - 66 , was observed in 1979 ( mazets , golenetskii , and guryan 1979 ) for the first time and became active again in 1992 ( kouveliotou _ et al . _ 1993 ) and also in 1998 ( kouveliotou _ et al . _ 1998 ; hurley _ et al_. 1999 ) . recently , the intense @xmath2-ray flare from sgr 1806 - 20 occured on december 27 , 2004 was reported ( palmer _ et al_. 2005 ; hurley _ et al_. 2005 ; mereghetti _ et al_. 2005 ) . ( 2005 ) have investigated the pulse profile and flare spectrum of sgr 1806 - 20 . their study may potentially give the information about the global field structure in the magnetosphere and may further promote theoretical magnetar models . their extream peak luminosity extends up to @xmath3 estimated at a distance of 10 kpc ( mazets _ et al_. 1999 ; feroci _ et al_. 2001 ) . typical x - ray luminosities of sgrs have been measured to be @xmath4-@xmath5 erg s@xmath6 ( hurley _ et al . _ 2000 ; thompson _ et al . _ 2000 ) , except with giant bursts sgrs 1900 + 14 and 1806 - 20 . shorter durations of initial flares of sgrs are comparable to the alfven crossing time of the core . the energy distribution shows good agreement with gutenberg - lichter law , which may indicate statistical similarity to earthquakes or solar flares ( cheng _ et al_. 1996 ; gogus _ et al_. 1999 , 2000 ) . the sgrs have spin periods in a small range of @xmath7-@xmath8 s with rapid spin down rate @xmath9 ss@xmath6 , which therefore give characteristic ages @xmath10 @xmath11 yr ( mazets _ et al_. 1979 ; kouveliotou _ et al_. 1998 ; hurley _ et al_. 1999a ) . on the other hand , the physical nature of axps is still uncertain due to poor observations , but there seem to be some similarities and differences between sgrs and axps . axps are energetic sources of pulsed x - ray emission , whose periods lie in a narrow range @xmath12-@xmath13 s , characteristic ages @xmath14 @xmath15 yr and x - ray luminosities @xmath16-@xmath5 erg s@xmath6(mereghetti 2000 ; thompson 2001 ) . axp sources likely have somewhat larger active ages and some of them have softer x - ray spectra compared with sgrs . one of the primary difference between them will be that axps so far have shown only quiescent x - ray emission with no bright active bursts such as giant flares . active ages of sgrs and axps are consistent with the observed evidence that these compact objects often give their location close to the edge of shell - type supernova remnants . neither sgrs nor axps show the presence of conspicuous counterparts at other wavelengths ( mereghetti & stella 1995 ; mereghetti _ et al_. 2002 ) . origin of these enourmous energetics is likely to come from their strong magnetic fields @xmath17-@xmath18 g estimated by their spin periods and spin down rates . only the magnetar model has been able to account for the enigmatic properties of a rare class of sgrs or axps . shear and alfvenic waves play an essential role on the energy transfer to the exterior in the burst - like phenomena observed in sgrs or giant flares . the alfven wave propagation in such a strongly magnetized star should be clarified theoretically . in our previous work , the propagation and transmission of torsional alfven waves have been studied , focused only on the fundamental mode of azimuthal wave number @xmath19 as a first step . in that work , the exterior of the star is examined qualitatively only , assuming two extreme cases : ( i ) corotating together with the star and ( ii ) static state independent of the stellar rotation . the former will come true when the plasma gas is trapped by closed magnetic field lines , and the latter will be realized without any other force . in any case , it might be inadequate at least in the following two points to place such strong constraints on the wave mode and on ambient circumstances of the star in the previous paper . first , under the more realistic situations various modes with azimuthal wave number @xmath20 are probably triggered by star quake . second , it is natural to assume that a scale height of the plasma gas is neither @xmath21 nor @xmath22 , but has a finite value of @xmath23 . et al_. ( 1998 ) pointed out that a relativistic _ fireball _ like those in classical grbs may exist in sgrs . recently , thompson and duncan ( 2001 ) have also suggested that after the initial hard spike emission , some lumps of hot plasma gas involving electron - positron pairs and high energy photons , that is , _ fireball , _ would be created on closed magnetic field lines . in fact , the fast decline and complete evaporation on x - ray light curve observed in the august 27 burst provides a clear evidence of the trapped _ fireball_. motivated by this , we thus extend our study to include some plasma gas spreading over the stellar surface . the main aim of this paper is to study the physical behavior of the torsional alfven waves with azimuthal wave number @xmath24 in the presence of corotating plasma confined within a certain finite distance @xmath23 and then to investigate the effects of these quantities @xmath24 and @xmath23 on the wave propagation and transmission . in section 2 , our model is self - consistently constructed and the relevant basic equations are formulated . these equations are almost the same as those derived in kojima and okita ( 2004 ) , but are summarized here for the paper to be self - contained . in section 3 , we shall derive the dispersion relation in a wkb way to show that the rotating plasma atmosphere plays a crucial role as a resonant cavity of the wave when a certain condition holds . in section 4 , transmission rates of the alfven waves are numerically calculated . in section 5 , their electromagnetic field structure is also discussed with the numerical results . in section 6 , we give a brief summary of our findings and their implications for the torsional alfven waves on rotating magnetars . both the magnetic field and the rotation of star lead to a complicated geometrical configuration . in this paper we assume some simplified conditions to understand the physical processes of the wave propagation . we here consider three - layered cylindrical model with radius @xmath25 . it is composed of the neutron star crust between @xmath26 and @xmath27 ( region ( 1 ) ) , corotating plasma above the stellar surface between @xmath27 and @xmath28 ( region ( 2 ) ) and static pure vacuum at @xmath29 ( region ( 3 ) ) as shown in fig.1 . the plasma gas filled in the atmosphere corotates with the crust at the same angular velocity @xmath30 . local magnetic fields permeate uniformly in each layer and point along the @xmath31 direction , @xmath32 , which represents open magnetic fields extending to infinity . alfven waves excited at the bottom of the crust @xmath33 cm owing to some mechanisms , travel upward along the local magnetic field lines . they are partially reflected and transmitted at the boundaries @xmath27 and @xmath34 , where the physical property of alfven waves significantly changes because of the effect of the background rotation , as shown below . this model , by setting @xmath35 , reduces to the case in which the plasma gas extends infinitely to the exterior of the star , while setting @xmath36 reduces to the simple case in which the exterior is filled with pure vacuum only . up to the present time , we have no observational information of the plasma size @xmath23 . in this work , @xmath23 is regarded as a free parameter in order to investigate its effect on the wave propagation and transmission . , region ( 2 ) corresponds to the _ fireball _ as a resonant cavity filled with plasma corotating with @xmath30 , and region ( 3 ) denotes a static pure vacuum . ] we shall now give some comments on the validity of this model by comparing the physical sizes @xmath37 , @xmath38 and @xmath23 with the stellar radius @xmath39 . in this model we have explicitly assumed that the local magnetic field has a @xmath31-component only . our model can be applied to the polar cap region , whose cylindrical radius @xmath37 is given by @xmath40 cm @xmath41 @xmath42 cm . curvature of the stellar surface and the magnetic field lines may be neglected within the polar cap region . in a similar way , it may be valid to assume that the local magnetic field lines are uniform if the thickness of each layer is smaller than the star radius , @xmath43 . we can also extend our model to the extreme case @xmath44 . even in this case , the validity of this model nevertheless holds good near the z - axis . in the remainder of this paper we ristrict our explicit calculations only to the axially symmetric small region within the polar cap radius , unless otherwise stated . in this section , we consider the propagation of torsional shear - alfven waves with various azimuthal modes on rotating magnetars . such waves are probably excited by the turbulent motion of the starquake in the deep crust , but above the neutron drip @xmath45 @xmath46 . many proposals for starquake model have been put forward ( e.g. , pacini and ruderman 1974 ) , but all of them are generally argued only for weakly magnetized neutron stars with @xmath47 g. therefore , their treatments are inadequate for magnetars , as they are . however , analogous mechanism may also occur in magnetars . in this paper , we do not discuss the triggering mechanism of the starquake itself , but focus only on the process by which electromagnetic shear waves , once generated , are propagated and transmitted from the deep crust , through a magnetized plasma , toward the vacuum exterior . we assume that the horizontal lagrange displacement @xmath48 is suddenly shaken in the deep interior at depth @xmath33 cm , nevertheless the matter remains immobile in the vertical direction due to strong gravity of the neutron star , @xmath49 cm s@xmath50 . therefore the following transverse wave condition may be easily satisfied.@xmath51as one of the simplest solutions , wave ansatz under this condition can be written formally as@xmath52here notation ` @xmath53 ' denotes the difference of helicity states . displacements and other perturbed quantities are always assumed to have a harmonic time dependency @xmath54 . the crust and the surrounding plasma can be regarded as a perfect conductor because ohmic dissipation timescales are very long compared with other timescales of interest . the frozen condition can read@xmath55background is assumed to rotate uniformly with @xmath56 , which means that the local electric field arround a highly conducting spherical star has only a radial component @xmath57 . the electric field now generates the goldreich - julian charge density@xmath58 transverse displacements of the crustal matter produce electromagnetic field perturbations propagating along the local magnetic field @xmath32 . a self - consistent set of maxwell s equations for perturbed electromagnetic fields @xmath59 and @xmath60 are @xmath61where @xmath62 denotes eulerian perturbation . by taking the perturbation of equation ( [ ff ] ) and then combining with equation ( [ max3 ] ) , @xmath63 and @xmath60 are expressed as@xmath64by using the relation between the displacement and the velocity perturbation @xmath65 together with equation ( [ displacement ] ) , one obtains @xmath66 e^{-i\left ( \omega t\mp m\phi \right ) } , \label{de2 } \\ \delta \mathbf{b } & = b_{\text{o}}\left ( \dfrac{\varpi } { \varpi _ { \text{pc}}}% \right ) ^{m-1}\dfrac{d\xi _ { \pm } } { dz}\left ( \mathbf{e}_{\varpi } \pm i% \mathbf{e}_{\phi } \right ) e^{-i\left ( \omega t\mp m\phi \right ) } . \label{db2}\end{aligned}\ ] ] equations ( [ de2 ] ) and ( [ db2 ] ) imply that if stellar rotation can be completely ignored , then @xmath59 , @xmath60 and the propagation vector @xmath67 form a mutually orthogonal set of vectors , say , tem modes . however , in the presence of rotation , such an orthogonality breaks down and thus longitudinal modes of the perturbed electric fields and currents are excited ( tm modes ) . other perturbed quantities related to the displacement are calculated by using above the results ( [ de2 ] ) and ( [ db2 ] ) @xmath68 \notag \\ & \times e^{-i\left ( \omega t\mp m\phi \right ) } , \\ \delta \rho _ { \text{e } } & = \dfrac{1}{4\pi } \nabla \cdot \delta \mathbf{e } \notag \\ & = \dfrac{b_{\text{o}}\omega \omega } { 4\pi c}\left ( \dfrac{\varpi } { \varpi _ { % \text{pc}}}\right ) ^{m-1}\dfrac{d^{2}\xi _ { \pm } } { dz^{2}}e^{-i\left ( \omega t\mp m\phi \right ) } .\end{aligned}\ ] ] the linearized equation of motion for uniformly rotating background from a static equilibrium state governing the torsional alfven waves is now given by@xmath69 \notag \\ & = & \delta s_{i}+\delta f_{i}+\left ( \mathbf{g}\delta \rho \right ) _ { i}-\left ( \nabla \delta p\right ) _ { i}\quad ( i=\varpi , \phi , z ) . \label{motion}\end{aligned}\]]profile of the mass density @xmath70 in the crust is obtained by solving the equation of state for degenerate electrons as follows ( blaes _ et al . _ 1989)@xmath71 ^{3/2}\text{g cm}^{-3}\text { , } \label{mass}\end{aligned}\]]where @xmath72 is an atomic mass unit , @xmath73 is the mean molecular weight per electron and others have usual physical meanings . the first term of rhs in equation ( [ motion ] ) , @xmath74 , is a perturbed elastic stress tensor associated with distorted matter in the crust and can be written in terms of @xmath75@xmath76 + \nabla _ { j}\left [ \mu \left ( \dfrac{\partial \xi ^{i}}{\partial x_{j}}+\dfrac{\partial \xi ^{j}}{\partial x_{i}}\right ) \right ] \notag \\ \quad ( i , j & = \varpi , \phi , z ) , \label{stress}\end{aligned}\]]where @xmath77 is a bulk modulus , @xmath78 is a shear modulus given by@xmath79with the ion number density @xmath80 and @xmath81 kronecker s delta . in the following calculations , @xmath82 will be adopted as a typical value in the crust . if one considers the complete transverse oscillation mode , the first term of equation ( [ stress ] ) vanishes owing to @xmath83 . the second term on rhs of equation ( [ motion ] ) can be resolved into perturbations of lorentz and coulomb forces@xmath84here we used equation ( [ de1 ] ) and unperturbed current @xmath85 induced by the background rotation . we further dropped the term @xmath86 , which is small near the z - axis or within the actual stellar radius @xmath39 . eliminating @xmath87 with the help of equation ( max4 ) , the electromagnetic force reduces to @xmath88the first term in equation ( [ em force ] ) implies the tension of the perturbed magnetic field , while the second one shows the magnetic pressure generated by the distorted matter . the last two terms of equation ( [ motion ] ) represent gravitational force @xmath89 and pressure gradient @xmath90 , respectively . however , unless one takes the p- or f - mode such as compressional waves and/or sonic waves into consideration , one can ignore them for simplicity in this model . we shall now derive the wave equation in region ( 1 ) shown in fig.1 . substituting equations ( [ stress ] ) and ( @xmath91 ) into equation of motion ( [ motion ] ) , one can obtain the following differential wave equation in terms of the displacement @xmath92 @xmath93 \xi _ { \pm } ^{(1)}=0 , \label{wave eq 1}\]]where @xmath94 and @xmath95 denote the effective shear modulus and the effective mass density defined as@xmath96the ratio of these quantities gives the shear - alfven wave speed in the crust @xmath97 in the above expression , the frequency @xmath98 measured in corotating frame for each helicity state has been introduced . we can limit this frequency to the positive regime @xmath99 for the symmetry @xmath100 . dimensionless function @xmath101 in equation ( [ wave eq 1 ] ) is formally defined as @xmath102which has a great influence on the dispersion relation of the wave especially with large @xmath24 not only in the inner surface , but also in the rotating plasma cloud , as discussed in the following section . note that if one considers the static background @xmath103 , wave equation ( [ wave eq 1 ] ) coincides with the one already derived by blaes _ et al_. ( 1989 ) . we now look for a wkb solution . in the deep interior the solution of wave equation ( [ wave eq 1 ] ) can be well asymptotically given by@xmath104 \right . \notag \\ & & \left . + b_{\pm } \exp \left [ i(\psi _ { \pm } ( z)-\omega t)\right ] \right\ } , \label{asy solution}\end{aligned}\]]with @xmath105 here @xmath106 denotes eikonals defined as@xmath107 } } { \widetilde{v } } , \label{eikonal}\]]for each mode . the first and second terms in equation ( [ asy solution ] ) represent the upward - propagating alfven wave with a complex incident amplitude @xmath108 and a downward - propagating wave with a complex reflection amplitude @xmath109 bounced at the stellar surface , respectively . we now turn to the wave behavior in region ( 2 ) @xmath110 mass density in this region is so small that one can formally take the limit @xmath111 . thus equation ( [ wave eq 1 ] ) reduces to @xmath112 \xi _ { \pm } ^{(2)}=0 . \label{wave eq 2}\]]the solution of this equation can be analytically written as@xmath113 + d_{\pm } \exp \left [ -i(k_{\pm } ^{(2)}z+\omega t)\right ] , \label{sol2}\]]where the wave number @xmath114 with each mode in the plasma is defined as@xmath115 } . \label{wave number 2}\ ] ] in region ( 3 ) with pure vacuum @xmath116 the wave equation becomes@xmath117owing to the absence of rotating matter , two helical states satisfy the same equation . we here dropped the notation ` @xmath53 ' . the solution is easily given by@xmath118 . \label{sol3}\]]the wave number thus takes an ordinal form as @xmath119 in the above expressions , @xmath120 and @xmath121 are complex incidence , reflection and transmission amplitudes in each region , respectively . these wave amplitudes and the wave numbers determine the transmission rate of the wave based on some boundary conditions . mathematical treatment will be given in section 2.4 . the physical property of alfven waves changes at the bottom and at the top of the rotating plasma layer , depending on the azimuthal wave number and the angular velocity of the background . we now require boundary conditions in the usual way in order to connect each wave solution continuously . from equations ( [ sol2 ] ) and ( [ sol3 ] ) , the continuity at the upper surface of the plasma , @xmath122 gives@xmath123 , \label{coeffi c } \\ d_{\pm } & = & e\dfrac{k_{\pm } ^{(2)}-k^{(3)}}{2k_{\pm } ^{(2)}}\exp \left [ i\left ( k_{\pm } ^{(2)}+k^{(3)}\right ) l\right ] . \label{coeffi d}\end{aligned}\]]substituting equations ( [ coeffi c ] ) and ( [ coeffi d ] ) into equation ( [ sol2 ] ) and then taking the derivative at the stellar surface , one obtains@xmath124where @xmath2 is a modification quantity due to background rotation of the surrounding plasma defined as @xmath125 -\varkappa _ { -}\exp \left [ i\varkappa _ { + } l\right ] } { \varkappa _ { + } \exp \left [ -i\varkappa _ { -}l\right ] + \varkappa _ { -}\exp \left [ i\varkappa _ { + } l\right ] } % , \]]with@xmath126detailed treatment and physical meaning of the @xmath2 are given in the appendix . in order to solve wave equation ( [ wave eq 1 ] ) , we consider the complex linear combination @xmath127 with specific solutions @xmath128 and @xmath129 to be satisfied with equation ( [ bc at surf ] ) . at the ends , one can obtain the explicit boundary condition at the stellar surface @xmath130 @xmath131logarithmic derivative of asymptotic solution has to be equal to that of numerical solution in the deep interior @xmath132 , so that we request@xmath133this yields@xmath134 .\]]finally , the reflection @xmath135 and transmission coefficients @xmath136 of the waves with each helicity propagating from the crust through the plasma toward the vacuum exterior are respectively expressed as@xmath137@xmath138 in this section , behavior of the torsional alfven waves is discussed based on the dispersion relations . eikonal equation ( [ eikonal ] ) depends on the depth through the functions @xmath101 and @xmath139 . for strong magnetic fields @xmath140 g considered in this paper , it can be shown that @xmath141 in equation ( [ eikonal ] ) varies weakly with the depth except for the inner region close to the star surface . the local dispersion relation can be approximately written as @xmath142 } . \label{dispersion relation}\]]this relation can also be obtained by taking a short wavelength limit , namely , high frequency limit . the dimensionless function @xmath101 involved in equation ( [ dispersion relation ] ) varies from @xmath143 in the deep interior to @xmath144 in the surface and exterior . this function represents the ratio between the rest mass energy density and the effective magnetic energy density . for @xmath143 , the rest mass energy dominates over the magnetic energy . this case corresponds to the classical limit . on the other hand , strong magnetic energy @xmath140 g easily overwhelms the rest mass energy of the electrons or ions in the low density region , where @xmath145 . in this case , relativistic displacement current should be included in the analysis . in order to investigate the propagation property of the wave , it is usefull to define the phase velocity @xmath146 in a straightforward manner as @xmath147which is also related to the refraction index @xmath148 for each mode . note that @xmath149 is not necessarily positive . it is well - known in plasma physics ( e.g. , wolfgang and rudolf 1996 ) that waves are in general reflected at the cut - off points where @xmath150 , and are absorbed at the resonant absorption points where @xmath151 . since the frequency @xmath152 in the rotating frame is here confined within the positive regime @xmath99 , @xmath153 is always real , that is , neither cut - off nor resonant absorption points appear in the positive helicity . on the other hand , there are both absorption and cut - off points for the negative mode . in the remainder of this section , our attention will be thus paid only to this mode for physical interests . the dispersion relations with negative mode are schematically shown for classical limit in fig.2 and for relativistic limit in fig.3 , respectively . in both diagrams the region @xmath154 stands for the non - propagation , viz . evanescent zone . vertical dotted line denotes the cut - off frequency of the wave . \(i ) classical limit as seen in fig.2 , evanescent zone appears only in the low frequency and narrow band width @xmath155 . in this classical treatment , the cut - off of the wave appears at @xmath156 @xmath157 , which can be physically interpreted as the fact that coriolis force interrupts the wave propagation in the rotating frame . in a high frequency region @xmath158 , waves are almost capable of propagating except for @xmath159 , at which waves are strongly absorbed by resonance with a rotating background . this is because refractivity diverges in this particular frequency . one notices that such a resonant absorption frequency @xmath160 in our model is formally analogous to the electron - cyclotron resonant frequency in the standard plasma physics ( wolfgang and rudolf 1996 ) . \(ii ) relativistic limit it is important to understand the new effect that appears in the relativistic case . figure 3 demonstrates that evanescent zone prevails in the frequency regime @xmath161 , whose band width is therefore broadened by large @xmath24 modes . cut - off frequency is given by @xmath162 . it is only a high frequency wave with @xmath163 that can always propagate in a wkb sense . by contrast with the classical limit , resonant absorption @xmath164 does not appear in this propagation regime , but in the evanescent regime . in a classical limit , @xmath165 . the region below horizontal axis corresponds to the evanescent zone . vertical dotted line denotes the cut - off frequency @xmath166.the resonant absorption frequency is given by @xmath167 . ] . the cut - off frequency is @xmath168 . the resonant absorption occurs within an evanescent regime at @xmath169 relation for negative helicity waves with @xmath170 in a classical limit , @xmath165 . the region below horizontal axis corresponds to the evanescent zone . vertical dotted line denotes the cut - off frequency @xmath166 . the resonant absorption frequency is given by @xmath171 . ] of the wave propagating above the potential . ] as is obvious from fig.3 , evanescency can not be neglected in large @xmath24 mode . in this subsection , we turn to the investigation of the spatial structure in the evanescent zone for large @xmath24 . we now introduce a new dependent variable @xmath172 defined by @xmath173 for the amplitude of the negative mode with @xmath174 . exploiting some arithmetic algebra , wave equation ( @xmath175 ) is now rewritten into a sturm - liouville type differential equation @xmath176 \psi _ { -}=0 , \label{new wave eq}\]]where@xmath177the first two terms concerning @xmath94 in square brackets in equation ( [ new wave eq ] ) become very small for strong magnetic fields @xmath178 g , since @xmath94 can be almost regarded as a constant . one can compare equation ( [ new wave eq ] ) with the one - dimensional schrdinger equation for box - type potential in a stationary state . physical quantities @xmath179 and @xmath180 respectively correspond to _ wave energy _ and _ potential _ in a formal sense . strictly , potential @xmath180 depends on the wave energy @xmath179 through the frequency @xmath181 . however , @xmath182 is here treated as a constant value irrespective of the position , so that the following discussions are valid without loss of generality . schematic profile of the _ effective potential _ @xmath180 and the _ wave energy _ @xmath179 are given in fig.4 as a function of distance @xmath31 . we draw the curve in region ( 1 ) somewhat exaggeratedly . the _ potential _ varies with the position through the function @xmath101 and the background rotation @xmath30 . potential__height is determined by the coupled quantity @xmath183 and its width is given by the size @xmath23 of the corotating zone . this result means that the large wave number lifts up the _ potential _ barrier , only if the background stellar medium rotates . in other words , if surrounding matter is not dragged by the star , say , being in static state @xmath184 , then the _ potential _ never rises , even though some high azimuthal waves exist . whether or not the wave can propagate and transmit out depends on the _ wave energy _ , that is , wave frequency @xmath181 . our argument deserves to be specially emphasized in two explicit regimes ; the evanescent frequency mode @xmath185 and the propagation mode @xmath186 . as is apparent in fig.4 , such low frequency waves excited in the interior can not help striking the potential barrier . since refractivity is zero on the critical curve @xmath188 , outgoing waves with @xmath189 are generally reflected , when they reach the potential wall . in this case perturbation of the electromagnetic fields can propagate as a wave . the present context in our model is almost concerned with the wkb frequency range . the propagation sometimes exhibits a remarkable property , if certain conditions are satisfied . as long as the wave frequency is much greater than the _ potential _ strength @xmath191 , the _ potential _ itself does not have much influence on the wave behavior . however , if the frequency becomes commensurable to the _ potential _ height @xmath192 , the existence of a _ potential _ barrier can not be ignored . especially if the wavelength is comparable to the _ potential _ width , i.e. , @xmath193 , then the waves will interfere with the _ potential _ barrier and their behavior will be strongly altered . this inherent property is expected to be more evident near the threshold frequency @xmath194 . motivated by this general consideration , we explicitly calculate the transmission rate in the next section . the numerical parameters are adopted to satisfy the above condition , @xmath195 cm and @xmath196 . using these parameters , we will examine how and to what extent the propagation and transmission are affected by the _ potential _ barrier . we numerically calculate the transmission rates ( [ trans rate ] ) of the wave propagating from the deep crust , through the corotating plasma envelope , toward the vacuum exterior subject to matching conditions at each boundaries . as already mentioned above , only the negative helicity wave is intriguing for physical interests . evanescent modes in a low frequency regime @xmath197 are excluded from this calculation . this constraint on wave frequency thereby guarantees that all waves are capable of propagating in a wkb sense . one can roughly estimate a typical wave frequency @xmath198 measured in the inertial frame by approximating as @xmath199 . appropriate frequency thus lies in a finite range @xmath200 s@xmath6 . the objective of this section is to explore the effects of azimuthal wave number @xmath24 and the size @xmath23 of the corotating plasma medium on the wave transmission . we here consider two kinds of specific torsional waves whose azimuthal number is ; ( i ) the fundamental , @xmath19 and ( ii ) much larger than one , @xmath201 . in each case , we further consider two apparently different circumstances in the exterior @xmath202 cm and in the absence of plasma @xmath21 cm . figures 5 and 6 respectively demonstrate the transmission coefficients for the negative helicity waves with two extreme cases ( i ) @xmath19 and ( ii ) @xmath203 as a function of the wave frequency measured in the inertial frame when @xmath204 g , typical for magnetars . in both figures , solid and dotted lines denote the results of @xmath205 and @xmath21 cm , respectively . the surrounding plasma is here assumed to corotate with the same angular velocity as that of the star , @xmath206 s@xmath6 . in each case , we got the following results . \(i ) the fundamental wave number : @xmath19 as seen in fig.5 , the transmission curve in the case of @xmath205 cm slightly shows a _ wiggling _ behavior . this curve intersects that of @xmath207 at @xmath208 s@xmath6 . however , the difference between both cases becomes undistinguishable in the high frequency regime @xmath209 s@xmath6 , since the wave frequency is much higher than the critical one @xmath210 s@xmath6 , which appeared as the potential barrier in section 3 . as the wave frequency gets higher , the transmission rate approaches unity asymptotically . since the overall property does not depend on @xmath23 , the potential barrier due to rotating plasma does not make any significant influence on the wave transmission for small azimuthal number @xmath211 . \(ii ) highly azimuthal wave number : @xmath212 in this high @xmath24 mode , we can find some suprising results . figure 6 shows that the transmission rate of @xmath205 cm is drastically enhanced at some selected frequencies . more importantly , such enhancements occur periodically at specific frequencies . at the top of the first wing @xmath213 @xmath214 s@xmath6 , the rate reaches the maximum @xmath215=@xmath216 , which corresponds to approximately @xmath217 times larger than that of the first bottom @xmath218 s@xmath6 . comparing with @xmath21 case at the same frequency @xmath219 , this maximum @xmath220 amounts to @xmath221 times larger . in this way , plasma effect is important at each top frequency , but unimportant at any bottom frequency . as shown in fig.6 , the transmission coefficient through the plasma layer is exactly equal to that of @xmath21 case at arbitrary wing bottom . in general rotating plasma has an effect in helping the escape of the waves except for the bottom frequencies . this point is different from that of case ( i ) . as a function of frequency when @xmath204 g and @xmath206 [email protected] and dotted lines denote the results of @xmath205and @xmath21 cm , respectively . ] . in this high @xmath24 mode transmission rate is highly enhanced at some selected frequencies due to the plasma envelope . ] in order to understand the transmission enhancement , we have also numerically computed the integral @xmath222 in some frequency bands for both @xmath205 and @xmath21 cm . we obtain some @xmath223 times enhancement in the narrow band @xmath224 and @xmath225 times in the broad band @xmath226 compared to the results of @xmath21 . such enhancement and periodic variation slowly decrease with increasing wave frequency . this property almost vanishes when the frequency becomes comparable to the critical frequency @xmath227 s@xmath6 . periodic enhancements in our numerical calculations can be interpreted as a consequence of wave interference within the rotating plasma cavity . in the following subsection , we investigate our results more quantitatively in a simplified analytical method . periodic enhancements on the wave transmission in our model can be satisfactorily rationalized by comparing with the wave propagation in homogeneous multi - media . non - relativistic alfven resonance itself in the magnetosphere around the earth and the sun have been widely recognized and discussed theoretically and observationally ( for recent reviews see leonovich and mazur 1997 ; waters 2000 ; li and wang 2001 ) . hollweg ( 1983 ) has theoretically studied wkb wave propagation in three homogeneous layers labeled by ( 1 ) , ( 2 ) and ( 3 ) separated at @xmath27 and @xmath34 as shown in fig.7 . in his work alfven waves are simply assumed to have constant wave numbers @xmath228 , @xmath229 and @xmath230 in each region . in this homogeneous model , the transmission coefficient @xmath231 can be analytically calculated by taking boundary conditions at the discontinuities @xmath27 and @xmath34 , @xmath232 ^{-1}. \label{trans2}\end{gathered}\ ] ] although the physical situation is different from our present work , similar expressions could be found also in our model . this formula ( [ trans2 ] ) implies that the transmission rate has a periodic structure depending on the wave number @xmath229 and the characteristic scale @xmath23 of the intermediate layer unless @xmath233 . extrema of the transmission rate work out to be @xmath234with @xmath235 . providing that the wave number in each region satisfies the inequalities @xmath236the coefficient of @xmath237 in equation ( [ trans2 ] ) becomes dominant compared to that of @xmath238 . in this limit , equation ( [ trans2 ] ) can be well approximated by@xmath239one notices that if @xmath240 , the transmission shows neither periodic variations nor enhancements . consequently , inequalities ( [ ineq k ] ) and ( [ ineq l ] ) give a set of resonant conditions of the wave . sterling and hollweg ( 1984 ) have subsequently considered a three layer model for solar flare , which is composed of a chromosphere , a spicule and a corona . in that work they have confidently suggested the possibility of alfvenic resonance on solar spicules and have shown a new aspect of spicules which may account for occasionally twisting motions of magnetic field lines , even when above resonant conditions approximately hold . conditions ( 46 ) and ( 47 ) may hold true also in our model except close to the thin regime beneath the stellar surface , whenever large twisted waves with @xmath20 propagate in the rotating background . we can now quantify the particular frequencies at which the wave resonance occur . by solving the quadratic equation @xmath241 , together with the periodic condition @xmath242 , the eigen frequencies of the negative mode are obtained as@xmath243with @xmath244 . this formula shows that the resonant frequencies depend critically on the azimuthal number @xmath24 and the plasma cavity length @xmath245 . they gently decrease with an increase in @xmath23 or @xmath24 . physically , this means that it takes longer time for the wave to go back and forth between the stellar surface and the top of the plasma layer and then this wave interferes with another one propagating from the crust into the cavity . in extremal cases , one finds @xmath246which coincides with the cut - off frequency in the classical limit . at the same time , transmission peaks become blended with neighboring resonance,@xmath247where @xmath248 . for @xmath206 s@xmath6 and @xmath205 cm , from equation ( [ resonant freq ] ) one can calculate some representative resonant frequencies ; the fundamental mode frequency @xmath249 s@xmath6 , the second mode @xmath250 s@xmath6 and the third @xmath251 s@xmath6 . our numerical work also gives similar results ; @xmath252 s@xmath6 , @xmath253 s@xmath6 and @xmath254 s@xmath6 , respectively . our results exhibit slightly positive deviations from the analytical ones . recall that the wave number @xmath255 in our magnetar model can not be regarded as a constant , since the mass density drastically changes in the vicinity of the stellar surface . this discrepancy would be probably attributed to the inhomogeneity in the crust . if a sharp boundary is formed at the stellar surface , such a difference may probably become small . from a comparison with the analytic model , we concluded that the transmission enhancements obtained in our model are thought to be a result of the alfven resonance on the rotating plasma cavity due to the potential barrier . and @xmath34 . if periodic boundary condition @xmath256 @xmath257 holds , intermediate layer acts as a resonant cavity and transmission is highly enhanced . ] in the preceding section we have elucidated that the transmission rates of wave are highly enhanced at some selected frequencies owing to the resonance effect in the rotating plasma cavity . resonance in the plasma portion has another spectacular nature . in this section , it will be shown that resonance has a great impact not only on the transmission rates , but also on the electromagnetic field structure associated with alfven waves . most of our applications will once again concern the wkb frequency regime of the negative helicity @xmath258 . hereafter we omit the subscript ` @xmath259 ' for simplicity . by substituting equation ( [ sol2 ] ) into ( [ de2 ] ) and ( [ db2 ] ) , we can explicitly have the expressions for electromagnetic field amplitudes in the plasma zone in terms of @xmath260 , normalized by a local magnetic field @xmath261@xmath262 , \end{gathered}\]]and@xmath263 , \end{gathered}\]]where @xmath264 is a radial profile given by @xmath265 @xmath266 and the other notations have the same meaning as the previous ones . the longitudinal component of the perturbed magnetic fields is always zero @xmath267 , because the matter has been assumed to be vertically immobile in the present work . on the contrary , only if the background rotates , the electric fields have a longitudinal component whose structure is essentially identical with that of the magnetic fields . thereby we only have to investigate the transverse components @xmath268 of the fields . for the sake of exploring the dependence of the azimuthal number @xmath24 on the field structure , we once more restrict our discussions within two kinds of extreme cases : ( i ) the fundamental mode @xmath19 and ( ii ) the large azimuthal wave number @xmath269 . in this limit , formulae ( 49 ) and ( 50 ) can be well approximated by@xmath270and@xmath271here we have dropped some small terms coupled to @xmath272 , since inequality @xmath273 holds for large @xmath24 . equations ( [ deruta b2 ] ) and ( [ deruta e2 ] ) have the implication that the fields are almost constant for small @xmath24 modes , but are sinusoidally changed with @xmath31 for large @xmath24 . in principle , the absolute value of the incident wave amplitude @xmath274 can not be determined in our linearized theory . if we are allowed to assume that @xmath275 , then the ratios of electromagnetic field amlitude with @xmath276 to those with @xmath19 are approximately given by@xmath277and @xmath278with @xmath279 . these quantities are much larger than unity because of the extra factor @xmath280 except for some special locations @xmath281 and @xmath282 @xmath283 which correspond to the nodes of the standing alfven wave . in the rotating plasma region normalized by that in the absence of plasma at some resonant frequencies in the first wing . the solid and the dotted lines denote the result at the top of wing @xmath284 s@xmath6 and at the bottom @xmath285 s@xmath6 , respectively . ] . field amplitudes are drastically and periodically enhanced for large @xmath24 . this enlargement is attributed to the resonance transmission and the conservation of energy flux as a multiplier effect . ] we will now examine the effect of the resonance in the rotating plasma cavity , whose thickness is given by @xmath286 cm , on the spatial structure of the perturbed electromagnetic fields . let us designate by @xmath287 the amplitude of the perturbed magnetic field within the plasmanormalized by that in the absence of the plasma @xmath288 .\text { \ \ } \label{alpha}\end{aligned}\]]the first term in square brackets is related to the deviation due to high torsional modes in the rotating background from the dispersion relation in a vacuum . the ratio @xmath289 is plotted against the height from the stellar surface for ( i ) @xmath19 in fig.8 and ( ii ) @xmath212 in fig.9 , respectively . solid lines denote the numerical results at the top of the wing @xmath214 s@xmath6 , while dotted lines denote the bottom @xmath290 s @xmath6 . the transmission rates at resonant frequencies obtained in section 3 are appropriately used in this calculation . \(i ) the fundamental wave number : @xmath19 as shown in fig.8 , in the case of @xmath291 the wave amplitude keeps one order of magnitude in the plasma . this result agrees well with the fact that the dispersion relation for small @xmath211 is almost equal to that of the electromagnetic fields in pure vacuum , @xmath292 . the background rotation therefore does not affect this small @xmath24 modes . \(ii ) highly azimuthal wave number : @xmath212 some significant differences can be found in this high mode . the perturbed magnetic field strength @xmath289 for @xmath212 at the resonant frequency @xmath293 is drastically amplified up to @xmath294 at @xmath295 cm . then the field strength @xmath289 is fallen down to @xmath296 at the top of the plasma layer , in which the first node of the standing alfven wave is formed . even at the bottom frequency @xmath297 the strength @xmath289 has about @xmath298 at @xmath299 cm . in the same way , the normalized amplitude of the perturbed electric field is expressed by@xmath300 , \ \label{beta}\end{aligned}\]]which is plotted in figs.10 and 11 . when @xmath212 , the field amplitude @xmath301 of @xmath219 has almost the same magnitude @xmath302 as @xmath287 near the stellar surface @xmath303 cm . but near the top of the plasma @xmath304 cm , in contrast to @xmath289 , @xmath305 has a maximum @xmath306 which corresponds to the loops of the standing wave . approximately , @xmath305 is much greater than @xmath289 because of the extra term @xmath307 in equation ( [ beta ] ) . such a field amplification can also be explained by considering a conservation of energy flux @xmath308 . the phase velocity @xmath309 of the mode @xmath20 in the plasma atmosphere is much smaller than that in the vacuum , as is confirmed in equations ( [ dispersion relation ] ) and ( [ phase velocity ] ) . assuming the same value @xmath310 at the stellar surface , which is irrelevant to @xmath23 , the amplitude @xmath311 is enhanced in the rotating plasma region so as to compensate for the slowing down of the wave propagation . in the rotating plasma at resonant frequencies in the first wing . solid line and dotted line respectively correspond to @xmath312 s@xmath6 and @xmath313 s@xmath6 . ] ] we have studied the propagation and transmission of torsional alfven waves along the rotation axis on magnetars by using three - layered cylindrical model . if the intermediate plasma layer rotates with angular velocity @xmath314 , the propagation property for largely twisted waves with @xmath315 is drastically modified . this middle zone can be regarded as a kind of a potential barrier , whose height is explicitly specified by @xmath316 . waves having an angular velocity of @xmath317 originates from coriolis force in the classical limit , while an additional quantity @xmath318 comes from the relativistic effect . this therefore means that the potential is lifted up higher in the large @xmath24 modes on the rotating background . it should be emphasized that this new finding is purely attributed to having incorporated the displacement current into the model . we have numerically computed the transmission rates of the torsional waves with large value @xmath20 driven in the crust , through the corotating plasma with a finite size @xmath44 , into the exterior . we found that the transmissions are strongly enhanced for large @xmath319 at selected wave frequencies satisfied with a periodic condition @xmath320 @xmath257 . such a transmission enhancement arises because the rotating plasma forms a resonant cavity which traps the wave energy by virtue of the strong reflections occuring at the transition region at the bottom and at the top of the plasma layer . efficient transmissions due to resonance is also compatible with the fact that the reflection of waves at the stellar surface can almost completely be extinguished . magnetospheric waveguide or resonance cavity can thus actually generate a set of coherent eigenmodes for high @xmath24 . resonance may be a signature of fundamental processes by which high order torsional oscillations of waves , if there exists , can transfer much more energy out of the magnetars especially in the low frequency regime . it has been well - known in quantum mechanics that when the potential width is comparable to ( or integer times ) de broglie wavelength of electrons @xmath321 , incident wave flux is bounded by finely tuning a phase relation with inversely propagating waves ( e.g. , schiff 1949 ) . then resonance is excited in the potential zone and the wave flux spends much of its time in the resonance cavity . similar resonance may occur also in our macroscopic model , if the characteristic scale of the evanescent zone is comparable to the wavelength of the alfven wave . we have investigated the spatial structure of the perturbed electromagnetic fields propagating in the rotating plasma , when standing alfven waves are formed . resonance in the plasma cavity for high @xmath24 modes has another spectacular aspect . at some points magnetic fields of the alfven waves are drastically amplified even up to a few @xmath302 times as large as those in the absence of plasma or in the static exterior . this amplification can be straightforwardly explained by both the transmission enhancement due to resonance and the conservation of wkb energy flux . in this way , once the alfven waves are resonantly excited and transmissions are greatly enhanced , a substantial part of the wave energy will probably be transferred into the ambient charged particles such as electrons or positrons confined within the fire balls associated with the bursts on magnetars . the charged particles entwined around magnetic field lines will then be violently swayed . these disturbances may eventually produce high energy @xmath2-ray or x - ray emissions from the magnetars . more importantly , if magnetic field lines at some altitudes are strongly distorted from equilibrium state by resonance and plasma fluid contracts into the surface , the field lines with antiparallel components will approach each other . such geometry of field lines will give rise to possible magnetic reconnection type events as thompson and duncan ( 2000 ) have suggested . the behavior of resonant mhd waves within the corona or spicules erupted from the solar surface have been studied so far . in fact , field line resonance has been found to occur on magnetic shells in the magnetosphere of the earth . but their physical treatment is inevitably restricted to the very weak magnetic fields . resonant property found in our magnetar model is very similar to that in solar astrophysics or planetary physics . we can expect that the same mechanism works also on the relativistic alfven waves even in a different environment , that is , in an extreme circumstance accompanied by very strong magnetic fields such as the magnetars . as a final remark , we shall address on the angular velocity @xmath30 of the star and the azimuthal wave number @xmath24 . in our model , the rotating background is throughout assumed to have a slow angular velocity @xmath206 s@xmath6 , which is actually observed in sgrs and axps . we concentrate our treatment on the torsional waves with a specific value @xmath212 as a possible limit of high @xmath24 . as shown in our work , the striking behavior of the resonance necessarily requires high order torsional modes , @xmath322 s@xmath323-@xmath324 . it is not clear whether or not such modes with high wave number @xmath24 realistically exist on the magnetars . however , one should keep it in mind that the local dispersion relation of the alfven waves are almost determined by the coupled quantity @xmath325 except for the high frequency regime . this means that the physical behavior of the wave with high @xmath326-@xmath324 on the slowly rotating background@xmath327 s@xmath6 is essentially equivalent to that of small @xmath19-@xmath11 and fast rotation @xmath328 s@xmath6 . neutron stars are often expected to have been a rapid rotator with @xmath329 s@xmath6 at a star - born period . our results would thus become much more effective especially on young magnetars . the physical effects of @xmath2 can be clarified by taking some explicit limits . specifically if one ignores the substantial thickness of the plasma layer @xmath330 , then the wave number of region ( 2 ) should be replaced with that of region ( 3 ) , @xmath331 . this yields that @xmath332 and @xmath333 . in this limit , the boundary condition reduces to @xmath334=ik^{(3)}$ ] , which clearly corresponds to the simple case that the exterior of the star is filled with pure vacuum or static plasma gas . while , if one imagines the very huge plasma gas corotating in the exterior and takes the limit formally @xmath335 , then the wave number of region ( 3 ) should be equal to that of region ( 2 ) , @xmath336 . we thus have that @xmath337 and @xmath338 , which yield the boundary condition @xmath339=ik_{\pm } ^{(2)}$ ] . one can draw some important facts from the boundary condition ( [ bc at surf ] ) . especially when the wave number satisfies the periodic condition @xmath340 @xmath257 , this yields @xmath341 and thereby the boundary condition for _ pure vacuum _ is recovered @xmath334=ik^{(3)}$ ] . the background rotation has no influence only on the alfven waves satisfied with this condition . whereas , when the wave number satisfies the another periodic condition @xmath342 @xmath257 , one obtains @xmath343 and thus the boundary condition works out to be @xmath334=ik^{(3)}(k_{\pm } ^{(2)}/k^{(3)})^{2}\gg ik^{(3)}$ ] . the rhs of this expression turns out that the waves satisfied with this condition are highly transmitted , as far as @xmath344 does not change much over one wavelength , say the validity of wkb approximation @xmath345 holds . this peculiar result means that the internal shear modes can be coupled onto the high frequency alfven modes at the stellar surface , which drastically enhances their transmission rate at certain wave numbers or frequencies .	torsional alfven waves are likely excited with bursts in rotating magnetars . these waves are probably propagated through corotating atmospheres toward a vacuum exterior . we have studied the physical effects of the azimuthal wave number and the characteristic height of the plasma medium on wave transmission . in this work , explicit calculations were carried out based on the three - layered cylindrical model . we found that the coupling strength between the internal shear and the external alfven modes is drastically enhanced , when resonance occurs in the corotating plasma cavity . the spatial structure of the electromagnetic fields in the resonance cavity is also investigated when alfven waves exhibit resonance . amma rays : bursts stars : neutron stars : magnetic fields
given a pseudo - anosov homeomorphism @xmath3 of a surface @xmath1 , let @xmath4 denote its dilatation . for any @xmath5 , we define @xmath6 it follows from work of penner @xcite that for @xmath7 sufficiently large , and for each closed surface @xmath8 of genus @xmath9 , there exists @xmath10 so that @xmath11 we refer to @xmath12 as the set of _ small dilatation pseudo - anosov homeomorphisms_. given a pseudo - anosov homeomorphism @xmath0 , let @xmath13 be the surface obtained by removing the singularities of the stable and unstable foliations for @xmath3 and let @xmath14 denote the restriction . the set of pseudo - anosov homeomorphisms @xmath15 is therefore also infinite . the main discovery contained in this paper is a universal finiteness phenomenon for all small dilatation pseudo - anosov homeomorphisms : they are , in an explicit sense described below , all `` generated '' by a finite number of examples . to give a first statement , let @xmath16 denote the homeomorphism classes of mapping tori of elements of @xmath17 . [ theorem : main1 ] the set @xmath16 is finite . we will begin by putting theorem [ theorem : main1 ] in context , and by explaining some of its restatements and corollaries . for a pseudo - anosov homeomorphism @xmath0 , the number @xmath4 gives a quantitative measure of several different dynamical properties of @xmath3 . for example , given any fixed metric on the surface , @xmath4 gives the growth rate of lengths of a geodesic under iteration by @xmath3 ( * ? ? ? * expos 9 , proposition 19 ) . the number @xmath18 gives the minimal topological entropy of any homeomorphism in the isotopy class of @xmath3 ( * ? ? ? * expos 10 , iv ) . moreover , a pseudo - anosov homeomorphism is essentially the unique minimizer in its isotopy class it is unique up to conjugacy by a homeomorphism isotopic to the identity ( * ? ? ? * expos 12 , thorme iii ) . from a more global perspective , recall that the set of isotopy classes of orientation preserving homeomorphisms @xmath19 forms a group called the mapping class group , denoted @xmath20 . this group acts properly discontinuously by isometries on the teichmller space @xmath21 with quotient the moduli space @xmath22 of riemann surfaces homeomorphic to @xmath1 . the closed geodesics in the orbifold @xmath22 correspond precisely to the conjugacy classes of mapping classes represented by pseudo - anosov homeomorphisms , and moreover , the length of a geodesic associated to a pseudo - anosov @xmath0 is @xmath18 . thus , the length spectrum of @xmath22 is the set @xmath23 arnoux yoccoz @xcite and ivanov @xcite proved that @xmath24 is a closed discrete subset of @xmath25 . it follows that @xmath24 has , for each @xmath1 , a least element , which we shall denote by @xmath26 . we can think of @xmath26 as the _ systole _ of @xmath22 . penner proved that there exists constants @xmath27 so that for all closed surfaces @xmath1 with @xmath28 , one has @xmath29 the proof of the lower bound comes from a spectral estimate for perron frobenius matrices , with @xmath30 ( see @xcite and @xcite ) . as such , this lower bound is valid for all surfaces @xmath1 with @xmath28 , including punctured surfaces . the upper bound is proven by constructing pseudo - anosov homeomorphisms @xmath31 on each closed surface of genus @xmath9 so that @xmath32 ; see also @xcite . the best known upper bound for @xmath33 is due to hironaka kin @xcite and independently minakawa @xcite , and is @xmath34 . the situation for punctured surfaces is more mysterious . there is a constant @xmath35 so that the upper bound of holds for punctured spheres and punctured tori ; see @xcite . however , tsai has shown that for a surface @xmath36 of fixed genus @xmath9 and variable number of punctures @xmath37 , there is no upper bound @xmath35 for @xmath38 , and in fact , this number grows like @xmath39 as @xmath37 tends to infinity ; see @xcite . one construction for small dilatation pseudo - anosov homeomorphisms is due to mcmullen @xcite . the construction , described in the next section , uses @xmath40manifolds and is the motivation for our results . given @xmath0 , the mapping torus @xmath41 is the total space of a fiber bundle @xmath42 , so that the fiber is the surface @xmath1 . if @xmath43 has dimension at least @xmath44 , then one can perturb the fibration to another fibration @xmath45 whose dual cohomology class is projectively close to the dual of @xmath42 ; see @xcite . in fact , work of thurston @xcite implies that there is an open cone in @xmath43 with the property that every integral class in this cone is dual to a fiber in a fibration . moreover , the absolute value of the euler characteristic of these fibers extends to a linear function on this cone which we denote @xmath46 . fried @xcite proved that the dilatation of the monodromy extends to a continuous function @xmath47 on this cone , such that the reciprocal of its logarithm is homogeneous . therefore , the product @xmath48 depends only on the projective class and varies continuously . mcmullen @xcite refined the analysis of the extension @xmath47 , producing a polynomial invariant whose roots determine @xmath47 , and which provides a much richer structure . in the course of his analysis , mcmullen observed that for a fixed @xmath40manifold @xmath41 , if @xmath49 is a sequence of distinct fibrations of @xmath50 with fiber @xmath51 and monodromy @xmath52 , for which the dual cohomology classes converge projectively to the dual of the original fibration @xmath42 , then @xmath53 , and @xmath54 in particular , @xmath55 is uniformly bounded , independently of @xmath56 , and therefore @xmath57 for some @xmath7 . there is a mild generalization of this construction obtained by replacing @xmath0 with @xmath14 ( as defined above ) and considering the resulting mapping torus @xmath58 . mcmullen s construction produces pseudo - anosov homeomorphisms of punctured surfaces from the monodromies of various fibrations of @xmath59 . by filling in any invariant set of punctures ( for example , all of them ) , we obtain homeomorphisms that may or may not be pseudo - anosov ( after filling in the punctures , there may be a @xmath60pronged singularity in the stable and unstable foliation , and this is not allowed for a pseudo - anosov ) . in any case , we can say that all of the pseudo - anosov homeomorphisms obtained in this way are _ generated by @xmath61_. theorem [ theorem : main1 ] can be restated as follows . [ corollary : fg ] given @xmath62 , there exists a finite set of pseudo - anosov homeomorphisms that generate all pseudo - anosov homeomorphisms @xmath63 in the sense above . corollary [ corollary : fg ] can be viewed as a kind of converse of mcmullen s construction . in view of corollary [ corollary : fg ] , we have the following natural question . for any given @xmath7 , can one explicitly find a finite set of pseudo - anosov homeomorphisms that generate @xmath12 ? puncturing the surface at the singularities is necessary for the validity of theorem [ theorem : main1 ] . indeed , the set of all pseudo - anosov homeomorphisms that occur as the monodromy for a fibration of a fixed @xmath40manifold have a uniform upper bound for the number of prongs at any singularity of the stable foliation . on the other hand , penner s original construction produces a sequence of pseudo - anosov homeomorphisms @xmath10 in which the number of prongs at a singularity tends to infinity with @xmath64 . so , the set @xmath65 of homeomorphism classes of mapping tori of elements of @xmath12 is an infinite set for @xmath7 sufficiently large . removing the singularities from the stable and unstable foliations of @xmath3 , then taking the mapping torus , is the same as _ drilling out _ the closed trajectories through the singular points of the suspension flow in @xmath66 . thus , we can reconstruct @xmath41 from @xmath58 by _ dehn filling _ ; see @xcite . the dehn - filled solid tori in @xmath50 are regular neighborhoods of the closed trajectories through the singular points and the _ dehn filling slopes _ are the ( minimal transverse ) intersections of @xmath1 with the boundaries of these neighborhoods . back in @xmath67 , the boundaries of the neighborhoods are tori that bound product neighborhoods of the ends of @xmath67 homeomorphic to @xmath68 . here the filling slopes are described as the intersections of @xmath69 with these tori and are called the _ boundary slopes of the fiber @xmath69_. moreover , since all of the manifolds in @xmath16 are mapping tori for pseudo - anosov homeomorphisms , they all admit a complete finite volume hyperbolic metric by thurston s geometrization theorem for fibered @xmath40manifolds ( see , e.g.@xcite , @xcite ) . as another corollary of theorem [ theorem : main1 ] , we obtain the following . [ corollary : maincor1 ] for each @xmath70 there exists @xmath71 with the following property . there are finitely many complete , noncompact , hyperbolic @xmath40manifolds @xmath72 fibering over @xmath73 , with the property that any @xmath74 occurs as the monodromy of some bundle obtained by dehn filling one of the @xmath75 along boundary slopes of a fiber . because hyperbolic volume decreases after dehn filling @xcite , as a corollary of corollary [ corollary : maincor1 ] , we have the following . [ c : bounded volume ] the set of hyperbolic volumes of @xmath40manifolds in @xmath65 is bounded by a constant depending only on @xmath7 . it follows that we can define a function @xmath76 by the formula : @xmath77 given a pseudo - anosov @xmath0 , we have @xmath63 for latexmath:[$\log p = we have the following . for any pseudo - anosov @xmath0 , we have @xmath79 using the relationship between hyperbolic volume and simplicial volume @xcite , a careful analysis of our proof of theorem [ theorem : main1 ] shows that @xmath80 is bounded above by an exponential function . for a homeomorphism @xmath3 of @xmath1 , let @xmath81 denote the translation length of @xmath3 , thought of as an isometry of @xmath21 with the weil petersson metric . brock @xcite has proven that the volume of @xmath66 and @xmath81 satisfy a bilipschitz relation ( see also agol @xcite ) , and in particular @xmath82 moreover , there is a relation between the weil petersson translation length and the teichmller translation length @xmath83 ( see @xcite ) , which implies @xmath84 however , brock s constant @xmath85 depends on the surface @xmath1 , and moreover @xmath86 when @xmath87 is sufficiently large . in particular , corollary [ c : bounded volume ] does not follow from these estimates . see also @xcite for a discussion relating volume to dilatation for a fixed surface . very little is known about the actual values of @xmath26 . it is easy to prove @xmath88 . the number @xmath26 is also known when @xmath87 is relatively small ; see @xcite . however , the exact value is not known for any surface of genus greater than @xmath44 ; see @xcite for some partial results for surfaces of genus less than @xmath89 . in spite of the fact that elements in the set @xmath90 seem very difficult to determine , the following corollary shows that infinitely many of these numbers are generated by a single example . [ c : magic ? ] there exists a complete , noncompact , finite volume , hyperbolic @xmath40manifold @xmath50 with the following property : there exist dehn fillings of @xmath50 giving an infinite sequence of fiberings over @xmath73 , with closed fibers @xmath91 of genus @xmath92 with @xmath93 , and with monodromy @xmath94 , so that @xmath95 we do not know of an explicit example of a hyperbolic @xmath40manifold as in corollary [ c : magic ? ] ; the work of @xcite gives one candidate . we now ask if the dehn filling is necessary in the statement of corollary [ c : magic ? ] : does there exist a single @xmath40manifold that contains infinitely many minimizers ? that is , does there exist a hyperbolic @xmath40manifold @xmath50 that fibers over the circle in infinitely many different ways @xmath96 , so that the monodromies @xmath97 satisfy @xmath98 ? to explain the motivation for the proof , we again consider mcmullen s construction . in a fixed fibered @xmath40manifold @xmath99 , oertel proved that all of the fibers @xmath51 for which the dual cohomology classes lie in the open cone described above are carried by a finite number of _ branched surfaces _ transverse to the suspension flow . a branched surface is a @xmath44dimensional analogue in a @xmath40manifold of a train track on a surface ; see @xcite . basically , an infinite sequence of fibers in @xmath50 whose dual cohomology classes are projectively converging are building up larger and larger `` product regions '' because they all live in a single @xmath40manifold . these product regions can be collapsed down , and the images of the fibers under this collapse define a finite number of branched surfaces . our proof follows this idea by trying to find `` large product regions '' in the mapping torus that can be collapsed , so that the fiber projects onto a branched surface of uniformly bounded complexity . in our case , we do not know that we are in a fixed @xmath40manifold ( indeed , we may not be ) , so the product regions must come from a different source than in the single @xmath40manifold setting ; this is where the small dilatation assumption is used . moreover , the collapsing construction we describe does not in general produce a branched surface in a @xmath40manifold . however , the failure occurs only along the closed trajectories through the singular points , and after removing the singularities , the result of collapsing is indeed a @xmath40manifold . while this is only a heuristic ( for example , the reader never actually needs to know what a branched surface is ) , it is helpful to keep in mind while reading the proof , which we now outline . first we replace all punctures by marked points , so @xmath1 is a closed surface with marked points . let @xmath41 be the mapping torus , which is now a compact @xmath40manifold . we associate to any small dilatation pseudo - anosov homeomorphism @xmath100 a markov partition @xmath101 with a `` small '' number of rectangles [ section [ s : markov partition ] ] . we prove that all but a uniformly bounded number of the rectangles of @xmath102 are essentially permuted [ lemma [ l : finite nonhomeo ] ] . this follows from the relationship with perron frobenius matrices and their adjacency graphs together with an application of a result of ham and song @xcite [ lemma [ l : matrix lemma ] ] . moreover , we prove that any rectangle meets a uniformly bounded number of other rectangles along its sides [ lemma [ l : bounded complexity 1 ] ] . these rectangles , and their images , are used to construct a cell structure @xmath103 on @xmath1 [ section [ s : cell structures ] ] . when a rectangle @xmath104 and all of the rectangles `` adjacent '' to it are taken homeomorphically onto other rectangles of the markov partition by both @xmath3 and @xmath105 , then we declare @xmath106 and @xmath107 to be `` @xmath103equivalent '' [ section [ s : equiv rel ] ] . this generates an equivalence relation on rectangles with a uniformly bounded number of equivalence classes [ corollary [ c : relation 1 prop ] ] . moreover , if two different rectangles are equivalent by a power of @xmath3 , then that power of @xmath3 is cellular on those rectangles with respect to the cell structure @xmath103 on @xmath1 [ proposition [ p : sim1 and y ] ] . in the mapping torus @xmath66 , the suspension flow @xmath108 applied to the rectangle for @xmath109 defines a `` box '' in the 3manifold , and applying this to all rectangles in @xmath101 produces a decomposition of @xmath50 into boxes which can be turned into a cell structure @xmath110 on @xmath50 so that @xmath111 with its cell structure @xmath103 is a subcomplex [ section [ s:3-mfd ] ] . if a rectangle @xmath106 is @xmath103equivalent to @xmath107 , then we call the box constructed from @xmath106 a `` filled box '' . the filled boxes stack end - to - end into `` prisms '' [ section [ s : quotient i ] ] . each of these prisms admits a product structure of an interval times a rectangle , via the suspension flow . we construct a quotient @xmath112 of @xmath50 by collapsing out the flow direction in each prism . the resulting @xmath40complex @xmath112 has a uniformly bounded number of cells in each dimension 0 , 1 , 2 , 3 , and with a uniform bound on the complexity of the attaching maps [ proposition [ p : hat w bounded ] ] . there are finitely many such @xmath40complexes [ proposition [ p : bounded complexity finite ] ] . the uniform boundedness is a result of careful analysis of the cell structures @xmath103 and @xmath110 on @xmath1 and @xmath50 , respectively . the flow lines through the singular and marked points are closed trajectories in @xmath50 that become 1subcomplexes in the quotient @xmath112 . we remove all of these closed trajectories from @xmath50 to produce @xmath113 and we remove their image @xmath60subcomplexes from @xmath112 to produce @xmath114 . we then prove that the associated quotient @xmath115 is homotopic to a homeomorphism [ theorem [ t : quotient htpc homeo ] ] . thus , @xmath116 is obtained from a finite list of @xmath40complexes by removing a finite @xmath60subcomplex , which completes the proof . * acknowledgements . * we would like to thank mladen bestvina , jeff brock , dick canary , and nathan dunfield for helpful conversations . our goal is to prove that the set of homeomorphism types of the @xmath40manifolds in @xmath16 is finite . we will accomplish this by finding a finite list of compact @xmath40dimensional cw complexes @xmath117 with the property that any @xmath40manifold in @xmath16 is obtained by removing a finite subcomplex of the @xmath60skeleton from one of the @xmath40complexes in @xmath117 . to prove that the set of @xmath40complexes in @xmath117 is finite , we will first find a constant @xmath118 so that each complex in @xmath117 is built from at most @xmath119 cells in each dimension . this alone is not enough to conclude finiteness . for example , one can construct infinitely many homeomorphism types of @xmath44complexes using one @xmath120cell , one @xmath60cell , and one @xmath44cell . to conclude finiteness , we must therefore also impose a bound to the complexity of our attaching maps . for an integer @xmath121 , we define the notion of _ @xmath119bounded complexity _ for an @xmath122cell of a cw complex @xmath123 in the following inductive manner . all @xmath120cells have @xmath119bounded complexity for all @xmath121 . having defined @xmath119bounded complexity for @xmath124cells , we say that an @xmath122cell has @xmath119bounded complexity if the attaching map @xmath125 has the following properties : 1 . the domain @xmath126 can be given the structure of an @xmath124complex with at most @xmath119 cells in each dimension so that each cell has @xmath119bounded complexity , and 2 . the attaching map is a homeomorphism on the interior of each cell . observe that a @xmath60cell has @xmath119bounded complexity for all @xmath127 . a @xmath44cell has @xmath119bounded complexity if the boundary is subdivided into at most @xmath119 arcs , and the attaching map sends the interior of these arcs homeomorphically onto the interiors of @xmath60cells in @xmath128 . a finite cell complex has _ @xmath119bounded complexity _ if each cell has @xmath119bounded complexity and there are at most @xmath119 cells in each dimension . such a cell complex is special case of a _ combinatorial complex _ ; see ( * ? ? ? * i.8 appendix ) . [ p : bounded complexity finite ] fix integers @xmath129 . the set of cw homeomorphism types of @xmath122complexes with @xmath119bounded complexity is finite . let @xmath123 be an @xmath122complex with @xmath119bounded complexity . we may subdivide each cell of @xmath123 in order to obtain a complex @xmath130 where each cell is a simplex and each attaching map is a homeomorphism on the interior of each cell ; in the language of @xcite , such a complex is called a @xmath131complex . what is more , we may choose @xmath130 so that the number of cells is bounded above by a constant that only depends on @xmath119 and @xmath122 . one way to construct @xmath130 from @xmath123 is to subdivide inductively by skeleta as follows . since any graph is a @xmath131complex , there is nothing to do for the 1-skeleton . to make the 2skeleton into a @xmath131complex , it suffices to add one vertex to the interior of each 2cell , and an edge connecting the new vertex to each vertex of the boundary of the 2cell ; in other words , we cone off the boundary of each 2cell . we then cone off the @xmath40cells , then @xmath132cells , and so on , until at last we cone off all the @xmath122cells . the number of simplices in any dimension is bounded by a function of @xmath119 and the dimension , so in particular , there is a uniformly bounded number of simplices , depending only on @xmath119 and @xmath122 . the second barycentric subdivision @xmath133 of @xmath130 is a simplicial complex ( this is true for any @xmath131complex ) , and the number of vertices of @xmath133 is bounded above by some constant @xmath134 that only depends on @xmath119 and @xmath122 . observe that @xmath123 is homeomorphic to @xmath133 . in fact , the cw homeomorphism type of @xmath123 is determined by the simplicial isomorphism type of @xmath133 , up to finite ambiguity : there are only finitely many cw homeomorphism types that subdivide to the simplicial isomorphism type of @xmath133 . we also note that @xmath133 is determined up to simplicial isomorphism by specifying which subsets of the @xmath134 vertices span a simplex . therefore , there are only finitely many simplicial isomorphism types of @xmath133 and so only finitely many cw homeomorphisms types of @xmath123 . our proof of theorem [ theorem : main1 ] requires a few facts about perron frobenius matrices and their associated adjacency graphs . after recalling these definitions , we give a result of ham and song @xcite that bounds the complexity of an adjacency graph in terms of the spectral radius of the associated matrix . ( 0,0)(0,0 ) ( 2.7,.8)@xmath60 ( 3.7,.8)@xmath44 ( 3.2,2.25)@xmath40 let @xmath135 be an @xmath136 _ perron frobenius matrix _ , that is , @xmath135 has nonnegative integer entries @xmath137 and there exists a positive integer @xmath56 so that each entry of @xmath138 is positive . associated to any perron frobenius matrix @xmath135 is its _ adjacency graph _ @xmath139 . this is a directed graph with @xmath122 vertices labeled @xmath140 and @xmath141 directed edges from vertex @xmath142 to vertex @xmath143 . for example , figure [ f : adjacency ] shows the adjacency graph for the matrix : @xmath144 the @xmath145entry of @xmath146 is the number of distinct combinatorial directed paths of length @xmath147 from vertex @xmath142 to vertex @xmath143 of @xmath139 . in particular , it follows that if @xmath135 is a perron frobenius matrix , then there is a directed path from any vertex to any other vertex . a perron frobenius matrix is sometimes defined as a matrix @xmath135 with nonnegative integer entries with the property that , for each pair @xmath145 , there is a @xmath56depending on @xmath145so that the @xmath145entry of @xmath138 is positive . as a consequence of our stronger definition , every positive power of a perron frobenius matrix is also perron frobenius . the perron frobenius theorem ( see e.g. @xcite ) tells us that @xmath135 has a positive eigenvalue @xmath148 , called the _ spectral radius _ , which is strictly greater than the modulus of every other eigenvalue and which has a @xmath60dimensional eigenspace spanned by a vector with all positive entries . moreover , one can bound @xmath148 from above and below by the maximal and minimal row sums , respectively : @xmath149 for a directed graph @xmath150 and a vertex @xmath151 of @xmath150 , let @xmath152 and @xmath153 denote the number of edges beginning and ending at @xmath151 , respectively . since each edge has exactly one initial endpoint and one terminal endpoint , it follows that the number of edges of @xmath150 is precisely @xmath154 the following fact is due to ham song ( * ? ? ? * lemma 3.1 ) [ l : matrix lemma ] let @xmath135 be an @xmath136 perron frobenius matrix with adjacency graph @xmath139 . we have @xmath155 both sums in the statement of lemma [ l : matrix lemma ] are equal to @xmath156 . in particular , lemma [ l : matrix lemma ] bounds the number of homeomorphism types of graphs @xmath139 in terms of @xmath148 , since @xmath139 can not have any valence @xmath60 vertices . as lemma [ l : matrix lemma ] is central to our paper , we give the proof due to ham and song . let @xmath157 be a spanning tree for @xmath139 . since @xmath158 , we have that @xmath157 has exactly @xmath122 vertices . since @xmath157 is a tree , we have @xmath159 , and so @xmath157 has @xmath160 edges . we claim that the number of directed paths of length @xmath122 starting from a vertex @xmath151 of @xmath139 is greater than or equal to the number of edges of @xmath139 not contained in @xmath157 . indeed , each path of length @xmath122 must leave @xmath157 , and so there is a surjective set map from the set of directed paths of length @xmath122 based at @xmath151 to the set of edges of @xmath139 not contained in @xmath157 : for each such path , take the first edge in the path that is not an edge of @xmath157 . let @xmath161 be the vertex of @xmath139 corresponding to the row of @xmath162 that has the smallest row sum . we have : @xmath163 since @xmath164 , we are done . let @xmath1 be a surface of genus @xmath64 with @xmath37 marked points ( marked points are more convenient for us than punctures in what follows ) . we still write @xmath165 and assume @xmath28 . first recall that one can describe a complex structure and integrable holomorphic quadratic differential @xmath166 on @xmath1 by a euclidean cone metric with the following properties : 1 . each cone angle has the form @xmath167 for some @xmath168 , with @xmath169 at any unmarked point @xmath170 . 2 . there is an orthogonal pair of singular foliations @xmath171 and @xmath172 on @xmath1 , called the _ horizontal _ and _ vertical foliations _ , respectively , with all singularities at the cone points , and with all leaves geodesic . such a metric has an atlas of _ preferred coordinates _ on the complement of the singularities , which are local isometries to @xmath173 and for which the vertical and horizontal foliations are sent to the vertical and horizontal foliations of @xmath173 . a homeomorphism @xmath0 is pseudo - anosov if and only if there exists a complex structure on @xmath1 , a quadratic differential @xmath166 on @xmath1 , and a @xmath174 , so that in any preferred coordinate chart for @xmath166 , the map @xmath3 is given by @xmath175 for some @xmath176 . in particular , observe that @xmath3 preserves @xmath171 and @xmath172 . the horizontal foliation is called the _ stable foliation _ , and the leaves are all stretched by @xmath177 ; the vertical foliation is called the _ unstable foliation _ , and the leaves of this foliation are contracted . the number @xmath177 is nothing other than the dilatation @xmath178 . the main technical tool we will use for studying pseudo - anosov homeomorphisms is the theory of markov partitions , which we now describe . let @xmath3 be a pseudo - anosov homeomorphism , and let @xmath166 be a quadratic differential as in our description of a pseudo - anosov homeomorphism . a _ rectangle _ ( for @xmath166 ) is an immersion @xmath179 \times [ 0,1 ] \to s\ ] ] that is affine with respect to the preferred coordinates , and that satisfies the following properties : 1 . @xmath180 maps the interior injectively onto an open set in @xmath1 , 2 . @xmath181)$ ] is contained in a leaf of @xmath172 and @xmath182 \times \{y\})$ ] is contained in a leaf of @xmath183 , for all @xmath184 and @xmath185 , and 3 . @xmath186 \times [ 0,1]))$ ] is a union of arcs of leaves and singularities of @xmath172 and @xmath171 . as is common practice , we abuse notation by confusing a rectangle and its image @xmath187 \times [ 0,1])$ ] . the _ interior _ of a rectangle @xmath188 is the @xmath180image of its interior . for any rectangle @xmath106 define @xmath189)\ \ \mbox{and}\ \ \partial_h r=\rho([0,1]\times \{0,1\}).\ ] ] we thus have @xmath190 . a _ markov partition _ for a pseudo - anosov homeomorphism @xmath0 is a finite set of rectangles @xmath191 satisfying the following : 1 . @xmath192 for @xmath193 , 2 . @xmath194 , 3 . @xmath195 , 4 . @xmath196 , and 5 . each marked point of @xmath1 lies on the boundary of some @xmath197 . we say that a markov partition @xmath198 for @xmath3 is _ small _ if @xmath199 the following lemma appears in the work of bestvina handel ; see 3.4 , 4.4 , and 5 of @xcite . suppose @xmath200 . every pseudo - anosov homeomorphism of @xmath1 admits a small markov partition . given a markov partition @xmath191 for @xmath0 , there is an associated nonnegative @xmath201 integral matrix @xmath202 , called the _ transition matrix _ for @xmath203 , whose @xmath145entry is @xmath204 where the absolute value sign denotes the number of components . that is , @xmath135 records how many times the @xmath205 rectangle `` maps over '' the @xmath206 rectangle after applying @xmath3 . the relation between the perron frobenius theory and the pseudo - anosov theory is provided by the following ; see e.g. ( * ? ? ? * expos 10 ) or @xcite . if @xmath0 is pseudo - anosov and @xmath102 is a markov partition for @xmath3 with transition matrix @xmath135 , then @xmath135 is a perron frobenius matrix and @xmath207 . let @xmath0 be a pseudo - anosov homeomorphism and let @xmath102 be a markov partition for @xmath3 . we say that a rectangle @xmath106 of @xmath102 is _ unmixed by @xmath3 _ if @xmath3 maps @xmath106 homeomorphically onto a rectangle @xmath208 , and we say that it is _ mixed by @xmath3 _ otherwise . given @xmath209 , any rectangle @xmath210 for which @xmath211 is called a _ target rectangle _ of @xmath106 for @xmath3 . the _ degree of @xmath209 for @xmath3 _ is defined as @xmath212 informally , @xmath213 records the total number of rectangles over which @xmath106 maps ( counted with multiplicity ) . the _ codegree _ of a rectangle @xmath209 for @xmath3 is the sum @xmath214 where the sum is taken over all rectangles @xmath210 having @xmath106 as a target rectangle . we will omit the dependence on @xmath3 when it is clear from context , and will simply write @xmath215 and @xmath216 . however , it will sometimes be important to make this distinction since a markov partition for @xmath3 is also a markov partition for every nontrivial power of @xmath3 . it is immediate from the definitions that if @xmath150 is the adjacency matrix associated to @xmath203 , and @xmath217 is the vertex associated to the rectangle @xmath209 , then @xmath218 and @xmath219 . translating other standard properties of adjacency graphs for perron frobenius matrices in terms of markov partitions , we obtain the following . [ p : graph / partition ] suppose @xmath0 is pseudo - anosov , @xmath102 is a markov partition for @xmath3 , and @xmath150 is the adjacency graph for the associated transition matrix . let @xmath209 , and let @xmath217 be the associated vertex of @xmath150 . the number of distinct combinatorial directed paths in @xmath150 that have length @xmath56 and that emanate from @xmath217 is precisely @xmath220 . the number of distinct combinatorial paths of length @xmath56 leading to @xmath217 is precisely @xmath221 . likewise , the following is immediate from the definitions and the related properties of perron frobenius matrices . [ p : static char ] let @xmath0 be a pseudo - anosov homeomorphism and let @xmath102 be a markov partition for @xmath3 . a rectangle is unmixed by @xmath3 if and only if @xmath222 and @xmath223 for the unique target rectangle @xmath224 of @xmath106 for @xmath3 . for a positive integer @xmath225 , if @xmath226 , then @xmath227 for all @xmath228 . similarly , for any @xmath225 , if @xmath229 for some target rectangle @xmath224 of @xmath106 by @xmath230 , then the same is true for some target rectangle of @xmath106 by @xmath231 for all @xmath228 . proposition [ p : static char ] immediately implies the following . [ c : persistent stretch ] suppose @xmath0 is pseudo - anosov and @xmath102 is a markov partition . then if @xmath209 is mixed by @xmath230 , then it is mixed by @xmath231 for all @xmath228 . we can also translate lemma [ l : matrix lemma ] into the context of markov partitions for small dilatation pseudo - anosov homeomorphisms . [ l : finite nonhomeo ] there is an integer @xmath232 , depending only on @xmath7 , so that if @xmath233 and if @xmath234 is a small markov partition for @xmath3 , then the number of rectangles of @xmath102 that are mixed by @xmath3 is at most @xmath235 . moreover , the sum of the degrees of the mixed rectangles is at most @xmath235 , and the sum of the codegrees of all targets of mixed rectangles is at most @xmath235 . by the definition of a small markov partition , we have @xmath236 , and so @xmath237 then , by the definition of @xmath12 , we have @xmath238 . let @xmath135 be the transition matrix for the pair @xmath203 , and let @xmath150 be the corresponding adjacency graph . we denote by @xmath217 the vertex of @xmath150 associated to a rectangle @xmath209 . since @xmath239 , lemma [ l : matrix lemma ] implies that @xmath240 from which it follows that @xmath150 has at most @xmath241 vertices @xmath217 with @xmath242 and at most @xmath241 vertices @xmath217 with @xmath243 . that is , @xmath222 for all but at most @xmath241 rectangles , and , among the rectangles @xmath106 with @xmath222 , all but at most @xmath241 of these have @xmath223 for their unique target rectangle @xmath224 . therefore , by proposition [ p : static char ] there are at most @xmath244 mixed rectangles . finally , observe that for any @xmath209 we have @xmath245 and similarly @xmath246 . since there are at most @xmath244 rectangles that are mixed , the sum of the degrees of the mixed rectangles is at most @xmath247 , and similarly the sums of the codegrees of target rectangles of mixed rectangles is at most @xmath247 . setting @xmath248 completes the proof . given a rectangle @xmath106 of a markov partition , we let @xmath249 denote its _ length _ , which is the length of either side of @xmath250 with respect to the cone metric associated to @xmath166 . similarly , we let @xmath251 denote its _ width _ , which is the length of either side of @xmath252 with respect to @xmath166 . [ l : bounded distortion ] let @xmath253 and let @xmath102 be a small markov partition for @xmath3 . if @xmath106 and @xmath224 are any two rectangles of @xmath102 , then @xmath254 it suffices to show that @xmath255 . by replacing @xmath3 with @xmath105 and switching the roles of @xmath106 and @xmath224 , the other inequalities will follow . let @xmath135 be the transition matrix for @xmath203 and let @xmath150 be its adjacency graph . for some @xmath256 is a target rectangle of @xmath197 for @xmath230 . that is , @xmath230 stretches @xmath197 over @xmath257 . in particular @xmath258 and so it suffices to show that @xmath259 . indeed : @xmath260 the equality uses the definition of a markov partition , the first inequality follows from the definition of @xmath12 , and the second inequality comes from the definition of a small markov partition . let @xmath3 be a pseudo - anosov homeomorphism of @xmath1 , and let @xmath234 be a markov partition for @xmath3 . the markov partition determines a cell structure on @xmath1 , which we denote by @xmath261 , as follows . the vertices of @xmath123 are the corners of the rectangles @xmath197 together with the marked points and the singular points of the stable ( and unstable ) foliation . the vertices divide the boundary of each rectangle of @xmath102 into a union of arcs , which we declare to be the @xmath60cells of our complex . finally , the 2cells are simply the rectangles themselves . we refer to @xmath60cells as either _ horizontal _ or _ vertical _ according to which foliation they lie in . for the next lemma , recall the notion of @xmath119bounded complexity , which is defined in section [ s : finiteness ] . [ l : bounded complexity 1 ] there is an integer @xmath262 so that for any @xmath63 , and for any small markov partition @xmath102 for @xmath3 , each of the cells of @xmath263 has @xmath264bounded complexity let @xmath209 . the first observation is that each of the four sides of @xmath106 contains at most one marked point or one singularity of the stable foliation for @xmath3 . indeed , if a side of @xmath265 were to contain more than one singularity or marked point , then the vertical or horizontal interval of @xmath265 connecting these two points would be a leaf of either the stable or unstable foliation . this is a contradiction since some power of @xmath3 would have to take this segment to itself , but @xmath3 nontrivially stretches or shrinks all leaves of the given foliation . now , for each @xmath60cell @xmath266 in the boundary of @xmath106 , at least one of the following holds : * @xmath266 is an entire edge of the boundary of a rectangle , or * @xmath266 has a corner of @xmath106 as an endpoint , or * @xmath266 has a marked point or singularity of the stable foliation for @xmath3 as an endpoint . along each side of @xmath106 there are at most two 1cells of type _ ( ii ) _ and , since there is at most one singularity or marked point , there are at most two 1cells of type _ ( iii)_. by lemma [ l : bounded distortion ] , the length of any rectangle of @xmath102 is at least @xmath267 and the width is at least @xmath268 . thus , there are at most @xmath241 @xmath60cells of type _ ( i ) _ along each side of @xmath106 . summarizing , there are at most @xmath269 @xmath60cells along each side of @xmath106 , and hence at most @xmath270 @xmath60cells in @xmath271 . thus , the constant @xmath272 satisfies the conclusion of the lemma ( this is trivial for @xmath120 and @xmath60cells ) . for any rectangle @xmath209 , its image @xmath107 is a union of subrectangles of rectangles in @xmath102 . thus , if we define @xmath273 then @xmath274 determines a rectangle decomposition of @xmath1 , with all rectangles being subrectangles of rectangles in @xmath102 . we define another cell structure @xmath275 on @xmath1 , in exactly the same way as above , using this new rectangle decomposition ; see figure [ f : markovsubdivide ] . ( 0,0)(0,0 ) ( 2.7,1.4)subdivide we can briefly explain the reason for introducing the cell structure @xmath103 . in section [ s:3-mfd ] , we will construct a cell structure on the mapping torus for @xmath3 . each 3cell will come from crossing a rectangle of @xmath102 with the unit interval . the gluing map for the one face of this `` box '' is given by @xmath3 , and hence would not be cellular in the @xmath123structure . the following gives bounds on the complexity of the cells of @xmath103 and the `` relative complexity '' of @xmath123 with respect to @xmath103 . [ l : bounded complexity 2 ] there is an integer @xmath276 so that if @xmath63 and @xmath277 is a small markov partition for @xmath3 , then the following hold . 1 . [ part1 ] each @xmath60cell of @xmath278 is a union of at most @xmath279 @xmath60cells of @xmath103 . [ part2 ] each @xmath44cell of @xmath123 is a union of at most @xmath279 @xmath44cells of @xmath103 . 3 . [ part3 ] for each @xmath44cell @xmath266 of @xmath123 , @xmath280 is a union of at most @xmath279 @xmath44cells of @xmath103 . [ part4 ] for each @xmath60cell @xmath266 of @xmath123 , @xmath280 is a union of at most @xmath279 @xmath60cells of @xmath103 . [ part5 ] each of the cells of @xmath281 has @xmath279bounded complexity . by the definition of a markov partition , the cell structure @xmath282 is obtained from the cell structure @xmath283 by subdividing each rectangle of @xmath123 into subrectangles along horizontal arcs . given @xmath284 , the number of rectangles in the subdivision is precisely @xmath216 , and is thus bounded by @xmath285 according to lemma [ l : finite nonhomeo ] . therefore , part [ part2 ] holds for @xmath286 . similarly , given any rectangle @xmath106 , @xmath107 is a union of @xmath287 @xmath44cells of @xmath103 , and so part [ part3 ] holds for @xmath286 . the subdivision of @xmath106 is obtained by first subdividing the vertical sides of @xmath106 by adding at most @xmath288 new @xmath120cells , then adding @xmath289 horizontal @xmath60cells from the left side of @xmath106 to the right . therefore , since each vertical @xmath60cell of @xmath123 lies in exactly two rectangles @xmath106 and @xmath224 , it is subdivided into at most @xmath290 @xmath60cells of @xmath103 by lemma [ l : finite nonhomeo ] . every horizontal @xmath60cell of @xmath123 is still a horizontal @xmath60cell of @xmath103 , so part [ part1 ] holds for @xmath291 . for part [ part5 ] , observe that any @xmath44cell @xmath266 of @xmath103 is a subrectangle of some rectangle @xmath209 . since each rectangle @xmath106 is a @xmath44cell of @xmath123 , it has at most @xmath264 @xmath60cells of @xmath123 in its boundary . by the previous paragraph , each of these @xmath60cells is subdivided into at most @xmath292 @xmath60cells of @xmath103 . therefore , there are at most @xmath293 @xmath60cells of @xmath103 in the boundary of @xmath106 , and thus no more than this number in the boundary of @xmath266 . part [ part5 ] holds for any @xmath294 ( this is trivial for @xmath120 and @xmath60cells ) . for part [ part4 ] , observe that if @xmath266 is a vertical @xmath60cell of @xmath123 , then @xmath280 is contained in the vertical boundary of some rectangle @xmath209 . as mentioned in the previous paragraph , the boundary of any rectangle has at most @xmath293 @xmath60cells of @xmath103 in its boundary , so @xmath280 is a union of at most @xmath293 @xmath60cells of @xmath103 . if @xmath266 is a horizontal @xmath60cell , then @xmath266 is contained in the horizontal boundary of some rectangle @xmath284 , and so @xmath280 is contained in the union of at most @xmath295 horizontal boundaries of rectangles of the subdivision . as already mentioned , the horizontal boundary of each of these rectangles is either one of the added @xmath60cells , or else a horizontal boundary of some rectangle of @xmath102 , which contains at most @xmath264 @xmath60cells of @xmath123 ( which are also the @xmath60cells of @xmath103 as they are contained in the horizontal edges of a rectangle ) . therefore , @xmath280 contains at most @xmath296 @xmath60cells of @xmath103 . therefore , part [ part4 ] holds for any @xmath297 . the lemma now follows by setting @xmath298 . given a pseudo - anosov @xmath253 , let @xmath102 be a small markov partition for @xmath3 . in this section we will use this data to construct a quotient space @xmath299 , and a cell structure @xmath300 on @xmath157 for which @xmath301 is cellular with respect to @xmath103 . moreover , we will prove that @xmath300 has uniformly bounded complexity . the quotient @xmath157 will be obtained by gluing together rectangles of @xmath102 via a certain equivalence relation . this relation has the property that if a rectangle @xmath106 is equivalent to @xmath224 , then there is a ( unique ) power @xmath302 that takes @xmath106 homeomorphically onto the rectangle @xmath224 . moreover , @xmath302 will be cellular with respect to @xmath103cell structures on @xmath106 and @xmath224 , respectively . this equivalence relation is most easily constructed from simpler equivalence relations . our approach will be to define a first approximation to the equivalence relation we are searching for , and then refine it ( twice ) to achieve the equivalence relation with the required properties . along the way we verify other properties that will be needed later . define an equivalence relation @xmath304 on @xmath102 by declaring @xmath305 if there exists @xmath306 so that @xmath307 takes @xmath106 homeomorphically onto @xmath224 . the ` h ' stands for `` homeomorphically '' . as no rectangle can be taken homeomorphically to itself by a nontrivial power of @xmath3 , it follows that if @xmath308 then there exists a unique integer @xmath309 for which @xmath310 . [ p : relation 1 prop ] let @xmath63 , and let @xmath102 be a small markov partition for @xmath3 . the @xmath303equivalence classes in @xmath102 have the form @xmath311 where each @xmath312 is an integer . there is a constant @xmath313 so that there are at most @xmath314 @xmath303equivalence classes . we first observe that @xmath308 if and only if @xmath106 is unmixed for @xmath231 and @xmath315 . proposition [ p : static char ] then implies that @xmath308 if and only if @xmath316 , @xmath224 is the unique target rectangle of @xmath106 and @xmath317 . corollary [ c : persistent stretch ] now implies that the @xmath303equivalence classes have the required form . to prove the second statement of the proposition , we observe that , according to the description of the @xmath303equivalence relation provided in the previous paragraph , the last rectangle @xmath106 in any given equivalence class ( which is well - defined by the first part of the proposition ) satisfies one of the following two conditions : * @xmath318 . * @xmath222 and @xmath319 , where @xmath224 is the unique target rectangle of @xmath106 . the number of rectangles @xmath106 of type _ ( i ) _ is bounded above by the constant @xmath320 from lemma [ l : finite nonhomeo ] . the number of type _ ( ii ) _ is bounded above by the sum of the codegrees of rectangles with codegree greater than 1 . again , by lemma [ l : finite nonhomeo ] , this is bounded by @xmath235 . thus , we may take @xmath314 to be @xmath292 . given @xmath253 and @xmath102 a small markov partition for @xmath3 , we index the rectangles of @xmath102 as @xmath321 so that 1 . the @xmath303equivalence classes all have the form @xmath322 . 2 . if @xmath323 is an equivalence class , then @xmath324 . that this is possible follows from proposition [ p : relation 1 prop ] . say that two rectangles @xmath325 are _ adjacent _ if @xmath326 contains at least one point that is not a singular point or a marked point . observe that , by definition , a rectangle is adjacent to itself . we let @xmath327 denote the set of rectangles adjacent to @xmath106 . we think of @xmath328 as the `` 1neighborhood '' of @xmath106 . [ l : rect in n1 ] there exists a constant @xmath329 so that if @xmath253 and @xmath102 is a small markov partition for @xmath3 , then the number of rectangles in @xmath328 is at most @xmath330 for any @xmath209 . the number of rectangles in @xmath328 is at most the number of @xmath60cells of @xmath123 in the boundary of @xmath106 plus @xmath331 . this is because each @xmath60cell contributes at most one new adjacent rectangle to @xmath328 , each corner vertex of @xmath106 contributes at most one more rectangle , and @xmath106 itself contributes @xmath60 . from lemma [ l : bounded complexity 2 ] it follows that we can take @xmath332 . now we define a refinement of @xmath303equivalence , called @xmath112equivalence , by declaring @xmath333 if @xmath302 does not mix @xmath334 for any @xmath335 ( the ` n ' stands for 1neighborhood ) . that is , @xmath302 maps the entire 1neighborhood @xmath328 to the 1neighborhood of @xmath224 , taking each rectangle homeomorphically onto another rectangle . we leave it to the reader to check that @xmath112equivalence is a refinement of @xmath303equivalence . figure [ f : equivalence ] shows the local picture of the surface after applying @xmath231 , for @xmath336 . we have labeled a few of the rectangles : @xmath337 @xmath338 the rectangles are related as follows : @xmath339 @xmath340 ( also @xmath341 , ) , and @xmath342 @xmath343 of course , each rectangle is equivalent to itself with respect to either relation . ( 0,0)(0,0 ) ( 2.45,4.65)@xmath197 ( 5.7,4.65)@xmath344 ( 2.5,2.5)@xmath345 ( 5.8,2.5)@xmath346 ( .5,4.5)@xmath1 ( 3.6,4.5)@xmath347 ( .35,2.35)@xmath348 ( 3.6,2.35)@xmath349 ( 2.4,3.2)@xmath257 ( 5.7,3.2)@xmath350 ( 2.45,1)@xmath351 ( 5.8,1)@xmath352 [ p : relation 2 prop ] let @xmath63 , and let @xmath102 be a small markov partition for @xmath3 . each @xmath112equivalence class has the form @xmath353 for some @xmath354 . there is a constant @xmath355 so that @xmath102 has at most @xmath356 @xmath112equivalence classes . similarly to the proof of proposition [ p : relation 1 prop ] , the first statement follows from corollary [ c : persistent stretch ] . now , consider an arbitrary @xmath303equivalence class : @xmath357 this is partitioned into its @xmath112equivalence classes as follows . the class divides at @xmath358 ( that is , @xmath359 begins a new @xmath112equivalence class ) if and only if @xmath360 contains a rectangle that is mixed by @xmath3 . by lemma [ l : finite nonhomeo ] , at most @xmath320 rectangles of @xmath102 are mixed . moreover , according to lemma [ l : rect in n1 ] , each rectangle in particular , each mixed rectangle is a @xmath60neighbor to at most @xmath361 rectangles . therefore , there are at most @xmath362 rectangles that are @xmath60neighbors of mixed rectangles . thus , each @xmath303equivalence class can be subdivided into at most @xmath363 @xmath112equivalence classes . according to proposition [ p : relation 1 prop ] the number of @xmath303equivalence classes is at most @xmath314 , and thus the proposition follows if we take @xmath364 . the next proposition explains one advantage of the @xmath112equivalence relation over the @xmath303equivalence relation . [ p : sim2 and x ] let @xmath253 and let @xmath102 be a small markov partition for @xmath3 . if @xmath365 , then @xmath366 is cellular with respect to @xmath123 . the cell structure @xmath123 is defined using the rectangles of @xmath102 , together with the way adjacent rectangles intersect one another , and the singular and marked points . since @xmath3 preserves the singular and marked points , and since @xmath365 implies @xmath302 maps all rectangles of @xmath328 homeomorphically onto a rectangle in @xmath367 , the result follows . we define a refinement of @xmath112equivalence , called @xmath103equivalence , by dividing each @xmath112equivalence class @xmath322 with more than one element into two @xmath103equivalence classes @xmath368 and @xmath369 . that is , we split off the initial element of each @xmath112equivalence class into its own @xmath103equivalence class . the @xmath103equivalence classes are also consecutive with respect to the indices and so we can refer to the initial and terminal rectangles of a @xmath103equivalence class . we write @xmath370 if @xmath106 and @xmath224 are @xmath103equivalent . the new feature of @xmath103equivalence is that any time two distinct rectangles are @xmath103equivalent , they both have codegree one . in fact , all rectangles in the @xmath60neighborhood are @xmath3images of unmixed rectangles . it follows that this equivalence relation behaves nicely with respect to the cell structure @xmath103 , hence the terminology ; see proposition [ p : sim1 and y ] below . if we take @xmath371 , we immediately obtain the following consequence of proposition [ p : relation 2 prop ] . [ c : relation 1 prop ] there is a constant @xmath372 so that for any @xmath63 , any small markov partition @xmath102 for @xmath3 has at most @xmath373 @xmath103equivalence classes . the next proposition is the analogue of proposition [ p : sim2 and x ] for @xmath103equivalence and the cell structure @xmath103 . [ p : sim1 and y ] let @xmath253 and let @xmath102 be a small markov partition for @xmath3 . if the @xmath103equivalence class of @xmath106 contains more than one element , and @xmath374 , then the @xmath123 and @xmath103cell structures on @xmath106 agree , and likewise for @xmath224 . moreover , @xmath366 is cellular with respect to @xmath103 . the assumption that the @xmath103equivalence class of @xmath106 contains more than one element is necessary for the first statement , since otherwise it would follow that the @xmath123 and @xmath103 cell structures coincide on all of @xmath1 . this would imply that @xmath3 is cellular with respect to @xmath123 , and hence finite order , which is absurd . since the @xmath103equivalence class of @xmath106 and @xmath224 contains more than one element , neither @xmath106 nor @xmath224 is the initial rectangle of the @xmath112equivalence class they lie in . thus , @xmath375 and @xmath376 are both elements of @xmath102 and @xmath377 by proposition [ p : sim2 and x ] , the maps @xmath378 are both cellular with respect to @xmath123 . thus , by the definition of @xmath103 , it follows that the @xmath103structures on @xmath106 and @xmath224 are exactly the same as the @xmath123structures on @xmath106 and @xmath224 , respectively . as @xmath365 , proposition [ p : sim2 and x ] guarantees that @xmath379 is cellular with respect to @xmath123 , and hence also with respect to @xmath103 , as required . we will use the @xmath103equivalence relation to construct a quotient of @xmath1 by gluing @xmath106 to @xmath224 by @xmath302 whenever @xmath370 . to better understand this quotient , we will study the equivalence relation on @xmath1 that this determines . this is most easily achieved by breaking the equivalence relation up into simpler relations as follows . we write @xmath380 to mean that the following conditions hold : 1 . @xmath381 2 . @xmath382 3 . @xmath370 4 . @xmath383 now write @xmath384 and say that @xmath385 _ is related to _ @xmath386 if @xmath380 for some @xmath387 . the relation @xmath388 is easily seen to be symmetric and reflexive , but it may not be transitive this is because if @xmath386 lies in two distinct rectangles @xmath224 and @xmath334 , it may be that @xmath380 and @xmath389 . we let @xmath390 denote the equivalence relation on @xmath1 obtained from the transitive closure of the relation @xmath388 on @xmath1 . [ l : hopping sim ] the @xmath390equivalence class of @xmath385 , which is contained in @xmath391 , consists of consecutive @xmath3-iterates of @xmath385 . what is more , if @xmath392 , then after possibly interchanging the roles of @xmath385 and @xmath386 , there exists @xmath393 , and a sequence of rectangles @xmath394 , so that @xmath395 with @xmath396 for @xmath397 . the second part of the lemma says that when @xmath392 , after possibly interchanging @xmath385 and @xmath386 , we can get from @xmath385 to @xmath386 moving forward through consecutive elements of the @xmath3orbit by applying the relation @xmath388 . we caution the reader that it may be necessary to interchange the roles of @xmath385 and @xmath386 , even when they lie in a periodic orbit . for a periodic point , the orbit @xmath391 is a finite set , in which case the ordering is a cyclic ordering . however , it still makes sense to say that a set consists of consecutive @xmath3-iterates of @xmath385 . note that if @xmath380 with @xmath398 then we also have @xmath399 for @xmath400 , since @xmath401 is contained in a single @xmath103equivalence class ( that is , the @xmath103equivalence classes of rectangles are consecutive ) . since @xmath390 is the transitive closure of @xmath388 , the full equivalence class is obtained by stringing together these sets whenever they intersect , and it follows that the equivalence class of @xmath385 consists of consecutive @xmath3-iterates of @xmath385 , and further that they are related as in statement of the lemma . the main purpose of this section is to describe the structure of the @xmath390equivalence classes and how they relate to the @xmath103 cell structure . given a point @xmath402 , set @xmath403 and @xmath404 that is , we look at all consecutive points in the orbit of @xmath385 that are equivalent , and take the infimum and supremum , respectively , of the consecutive exponents ( beginning at @xmath120 ) that occur . according to lemma [ l : hopping sim ] , the set @xmath405 is the @xmath390equivalence class of @xmath385 . observe that if @xmath385 is a fixed point of @xmath3 then @xmath406 and @xmath407 . [ l : orbit sim ] if @xmath385 is not a marked point or singular point , then @xmath408 . in particular , it can not be the case that the entire @xmath3orbit of @xmath385 lies in the same @xmath390equivalence class . first suppose that @xmath385 is a periodic point of order @xmath122 . the only way that @xmath409 or @xmath410 is if any two points of the orbit are @xmath390equivalent . up to changing @xmath385 within its orbit , we may assume that there is a rectangle @xmath106 containing @xmath385 that is mixed by @xmath3 ; indeed , otherwise the pseudo - anosov map @xmath411 would preserve the collection of rectangles containing @xmath385 for all @xmath56 which is impossible . we now prove @xmath412 , which will complete the proof in the case of a periodic point . the proof is by contradiction , so assume that @xmath413 . according to lemma [ l : hopping sim ] , there are two cases to consider depending on whether or not the roles of @xmath385 and @xmath414 must be interchanged . thus , we either have a rectangle @xmath224 for which @xmath415 or else there exists a sequence of rectangles @xmath416 so that @xmath417 for @xmath418 . in particular , in the second case we have @xmath419 . the first case is clearly impossible since @xmath224 is adjacent to @xmath106 which is mixed by @xmath3 , and hence @xmath420 . in particular , it follows that @xmath421 , which is a contradiction . in the second case , we have @xmath422 which implies @xmath423 . since @xmath424 is a rectangle containing @xmath385 , it is adjacent to ( the mixed rectangle ) @xmath106 and therefore @xmath425 , another contradiction . therefore , it must be the case that @xmath426 , as required . we now consider the case where @xmath385 is not a periodic point for @xmath3 . by lemma [ l : hopping sim ] , the equivalence class of @xmath385 consists of consecutive @xmath3-iterates . let @xmath143 be a positive integer so that one of the rectangles containing @xmath427 , say @xmath106 , is mixed by @xmath3 . as in the periodic case , it follows that @xmath428 . since @xmath385 is aperiodic , it follows from lemma [ l : hopping sim ] that @xmath429 , so @xmath430 . similarly , note that there is some @xmath431 so that @xmath432 is contained in a rectangle @xmath106 that is mixed by @xmath3 . so , @xmath433 , and again aperiodicity of @xmath385 together with lemma [ l : hopping sim ] implies @xmath434 , so @xmath435 . we now prove that @xmath436 depends only on the cell containing @xmath385 in its interior . [ p : structure of sim ] let @xmath266 be a cell of @xmath103 that is not a singular point or a marked point . if @xmath437 , then @xmath438 . if @xmath439 and @xmath440 , then @xmath441 in particular , we can define @xmath442 for any @xmath439 ( independently of the choice of @xmath443 ) , and for every integer @xmath444 $ ] , the map @xmath445 is cellular with respect to @xmath103 . we suppose first that @xmath439 and @xmath446 and we prove that @xmath447 . if @xmath448 , there is nothing to prove so suppose @xmath449 . combining lemmas [ l : hopping sim ] and [ l : orbit sim ] , there exists a sequence of rectangles @xmath450 , so that @xmath451 for @xmath397 . for each @xmath143 , proposition [ p : sim1 and y ] implies that @xmath452 is cellular with respect to @xmath103 . since @xmath266 is a cell in @xmath453 , by induction and proposition [ p : sim1 and y ] , we see that @xmath454 is a cell in both @xmath455 and @xmath456 . hence @xmath457 and @xmath458 , and therefore @xmath459 for every @xmath397 . it follows that @xmath460 for @xmath461 and thus @xmath462 . observe that if @xmath463 , we can reverse the roles of @xmath385 and @xmath464 to obtain @xmath465 , and hence @xmath466 . otherwise , @xmath440 , and we only know @xmath447 . the proof that @xmath467 if @xmath463 and @xmath468 if @xmath469 follows a similar argument . that we can define @xmath442 for any @xmath439 now follows . finally , the fact that @xmath445 is cellular is proven in the course of the proof above . denote @xmath470 by @xmath157 and let @xmath299 be the quotient map . [ p : w structure ] let @xmath253 and let @xmath102 be a small markov partition for @xmath3 . there is a cell structure @xmath471 on @xmath157 , so that @xmath301 is cellular with respect to the cell structure @xmath275 on @xmath1 . moreover , for each cell @xmath266 of @xmath103 , @xmath301 restricts to a homeomorphism from the interior of @xmath266 onto the interior of a cell @xmath472 of @xmath300 . finally , there is a @xmath473 so that @xmath300 has @xmath474bounded complexity . it follows from proposition [ p : structure of sim ] that @xmath390 defines an equivalence relation on the cells of @xmath103 , which we also call @xmath390 , by declaring @xmath475 if @xmath266 is a cell in @xmath103 that is neither a singular point nor a marked point , and @xmath476 , then appealing to lemma [ l : hopping sim ] , lemma [ l : orbit sim ] , and proposition [ p : structure of sim ] , there is a unique integer @xmath477 $ ] so that @xmath478 . moreover , by proposition [ p : structure of sim ] , @xmath479 is cellular with respect to @xmath103 . an exercise in cw topology shows that @xmath157 admits a cell structure @xmath300 with one cell for each equivalence class of cells in @xmath103 . moreover , the characteristic maps for these cells can be taken to be the characteristic maps for cells of @xmath103 , composed with @xmath301 . indeed , if @xmath476 in @xmath103 , and @xmath266 and @xmath480 are @xmath37cells with @xmath481 , then we may assume that the characteristic maps @xmath482 and @xmath483 for @xmath266 and @xmath480 , respectively , are related by @xmath484 of course , for @xmath120cells , the characteristic maps are canonical . it follows that @xmath301 is cellular and is a homeomorphism on the interior of any cell . since @xmath301 is cellular and is a homeomorphism on the interior of each cell , lemma [ l : bounded complexity 2 ] implies that each @xmath44cell of @xmath300 has @xmath474bounded complexity for any @xmath485 . recall that the @xmath120 and @xmath60cells trivially have @xmath474bounded complexity for all @xmath486 . lemma [ l : bounded complexity 2 ] also implies that every rectangle @xmath209 is subdivided into at most @xmath279 @xmath44cells of @xmath103 . by corollary [ c : relation 1 prop ] there are at most @xmath373 @xmath103equivalence classes of rectangles , and therefore there are at most @xmath487 @xmath44cells of @xmath300 . since each @xmath44cell has @xmath279bounded complexity , there are at most @xmath488 @xmath120 and @xmath60cells . thus , @xmath300 has @xmath474bounded complexity for @xmath489 . let @xmath0 be a pseudo - anosov homeomorphism , and let @xmath41 denote the mapping torus . this is the @xmath40manifold obtained as a quotient of @xmath490 $ ] by identifying @xmath491 with @xmath492 for all @xmath402 . we view @xmath1 as embedded in @xmath50 as follows : @xmath493 \to m.\ ] ] let @xmath494 denote the suspension flow on @xmath50 ; this is the flow on @xmath50 determined by the local flow @xmath495 on @xmath490 $ ] . the time - one map restricted to @xmath1 is the first return map , that is , @xmath496 . from this it follows that @xmath497 for every integer @xmath498 and @xmath402 . we also have the punctured surface version . the punctured surface @xmath69 embeds in @xmath1 and is @xmath3invariant . hence , we may view @xmath67 as @xmath499 , the mapping torus of @xmath500 , embedded in @xmath501 . let @xmath253 and let @xmath502 be a small markov partition for @xmath3 . for each rectangle @xmath197 , there is an associated _ box _ , defined by @xmath503 more precisely , if @xmath504 \times [ 0,1 ] \to r_i$ ] is the parameterized rectangle , then the box is parameterized as @xmath505 \times [ 0,1 ] ) \times [ 0,1 ] \to b_i\ ] ] where @xmath506 as with rectangles , the map @xmath507 is only an embedding when restricted to the interior . and as in the case of rectangles , we abuse notation , and refer to @xmath508 as a subset of @xmath50 , although it is formally a map into @xmath50 . we impose on @xmath50 a cell structure @xmath509 defined by the cell structures @xmath123 and @xmath103 on @xmath1 as follows . the @xmath120cells of @xmath110 are the @xmath120cells of @xmath103 ( recall that @xmath111 ) . the @xmath60cells of @xmath110 are the @xmath60cells of @xmath103 , called _ surface @xmath60cells _ , together with suspensions of @xmath120cells of @xmath123 : @xmath510 because @xmath511 and since @xmath512 , the boundary of each of these 1cells is contained in @xmath513 , as required . we call these the _ suspension @xmath60cells_. the @xmath44cells of @xmath110 are the @xmath44cells of @xmath103 , called _ surface @xmath44cells _ , together with suspensions of @xmath60cells of @xmath123 : @xmath514 observe that for each @xmath60cell @xmath266 of @xmath123 , both @xmath266 and @xmath280 are @xmath60subcomplexes of @xmath103 . furthermore , since the boundary of @xmath266 is contained in @xmath515 , we see that the boundary of the @xmath44cell defined by the above suspension is contained in @xmath516 , as required . we call these @xmath44cells _ suspension @xmath44cells_. finally , the @xmath40cells of @xmath110 are the boxes . all @xmath40cells can be thought of as a _ suspension @xmath40cells _ , since they are suspensions of surface @xmath44cells . [ p : hat y ] there exists a positive integer @xmath517 with the following property . if @xmath253 and if @xmath102 is a small markov partition for @xmath3 , then each cell of @xmath509 has @xmath518bounded complexity . all @xmath120cells and @xmath60cells trivially have @xmath518bounded complexity for all @xmath519 . since @xmath1 is embedded in @xmath50 as the subcomplex @xmath520 , it follows from part [ part5 ] of lemma [ l : bounded complexity 2 ] that the surface @xmath44cells of @xmath110 have @xmath518bounded complexity for any @xmath521 . if @xmath266 is a suspension @xmath44cell obtained by suspending a @xmath60cell @xmath522 of @xmath123 , then the boundary of @xmath266 consists of @xmath44 suspension @xmath60cells , and by parts [ part1 ] and [ part4 ] of lemma [ l : bounded complexity 2 ] , at most @xmath523 surface @xmath60cells . therefore , these @xmath44cells have @xmath518bounded complexity for any @xmath524 . the boundary of each @xmath40cell is a union 6 rectangles : a `` bottom '' and a `` top '' ( which are rectangles of @xmath102 and @xmath525 , respectively ) , and four `` suspension sides '' . the number of @xmath120cells is just the number of @xmath120cells in the top and bottom rectangles ( since all vertices lie in @xmath1 ) . each of the top and bottom rectangles is a union of at most @xmath279 surface @xmath44cells by parts [ part2 ] and [ part3 ] of lemma [ l : bounded complexity 2 ] . each surface @xmath44cell has @xmath279bounded complexity , so has at most @xmath279 vertices in its boundary . it follows that each @xmath40cell has at most @xmath526 vertices in its boundary . the number of suspension @xmath44cells in the boundary of a @xmath40cell @xmath508 is the number of @xmath60cells of @xmath123 in the boundary of @xmath527 . since the cells of @xmath123 have @xmath528bounded complexity by lemma [ l : bounded complexity 1 ] , and the @xmath44cells of @xmath123 are precisely the rectangles of @xmath102 , it follows that the boundary of a @xmath40cell has at most @xmath264 suspension @xmath44cells in its boundary . combining this with the bounds from the previous paragraph on the number of surface @xmath44cells , it follows that the boundary of a @xmath40cell has at most @xmath529 @xmath44cells in its boundary . finally , since the boundary of a @xmath40cell is a @xmath44sphere , the euler characteristic tells us that the number of @xmath60cells in the boundary is 2 less than the sum of the numbers of @xmath120cells and @xmath44cells , and so is at most @xmath530 . it follows that if @xmath531 , then each @xmath40cell has @xmath518bounded complexity . therefore , setting @xmath532 completes the proof . a subset of the suspension @xmath60cells are the _ singular _ and _ marked @xmath60cells _ ; these are the @xmath60cells of the form @xmath533 for @xmath151 either a singular point or marked point , respectively . these @xmath60cells , together with their vertices , form @xmath60dimensional subcomplexes called the _ singular _ and _ marked subcomplexes_. these subcomplexes are unions of circles in @xmath50 , and @xmath67 is obtained from @xmath50 by removing them . in this section we use the notion of @xmath103equivalence to produce a quotient of the compact @xmath40manifold @xmath41 ; the quotient we obtain will be compact , but might not be a manifold . we will prove that this quotient admits a @xmath40dimensional cell structure with uniformly bounded complexity for which the quotient map is cellular . the singular and marked subcomplexes of @xmath110 will define subcomplexes of the quotient , and in the next section we will prove that the space obtained by removing these subcomplexes from the quotient is homeomorphic to @xmath67 , the corresponding mapping torus of the punctured surface @xmath69 . first , we define and analyze the quotients . let @xmath253 be a pseudo - anosov homeomorphism and let @xmath191 be a small markov partition for @xmath3 , indexed so that each @xmath103equivalence class has the form @xmath534 with @xmath324 ; see section [ s : equiv rel ] . let @xmath108 be the suspension flow on @xmath50 . as in section [ s:3-mfd ] , we obtain collection of boxes @xmath535 . let @xmath110 be the cell structure on @xmath50 as described in section [ s:3-mfd ] . to each @xmath103equivalence class of rectangles , we will associate a _ prism _ as follows . if @xmath534 is a @xmath103equivalence class , then the associated prism is @xmath536 the rectangles @xmath197 and @xmath537 are the _ bottom _ and _ top _ of the prism , respectively . alternatively , if @xmath322 is a @xmath103equivalence class with more than one element , then the associated prism is @xmath538 on the other hand if @xmath368 is a @xmath103equivalence class consisting of the single rectangle @xmath197 , then @xmath539 . if a box is contained in a prism , then we will say it is a _ filled box_. see figure [ f : prism ] . let @xmath540 be the union of the prisms in @xmath50 . for clarification , we note that @xmath540 is the union of @xmath1 together with the set of filled boxes , and is a subcomplex of @xmath110 . the prisms define an equivalence relation @xmath541 on @xmath50 that is a `` continuous version '' of @xmath390 on @xmath1 . more precisely , we declare @xmath542 if and only if either @xmath543 or else there exists an interval @xmath544 $ ] or @xmath545 $ ] in @xmath546 so that @xmath547 and @xmath548 that is , @xmath542 if there is an arc of a flow line containing @xmath464 and @xmath549 that is contained in @xmath540 . we will be particularly interested in the flow lines in @xmath50 through points in @xmath1 . we denote the flow line in @xmath50 through @xmath402 by @xmath550 [ p : prism relation ] the equivalence relation @xmath541 restricted to @xmath1 is precisely the equivalence relation @xmath390 . according to lemma [ l : hopping sim ] , if @xmath392 , then after reversing the roles of @xmath385 and @xmath386 , we have rectangles @xmath551 so that @xmath552 with @xmath396 for @xmath397 . therefore , @xmath553 by the definition of the @xmath388 relation , we have that @xmath554 , and so @xmath555 is a filled box . thus @xmath556 for each @xmath397 . by transitivity , @xmath557 . for the other direction , suppose @xmath558 . we have that @xmath559 and the arc @xmath560 from @xmath385 to @xmath386 is contained in @xmath540 . say that @xmath561 for @xmath562 ( reverse the roles of @xmath385 and @xmath386 if necessary ) . we can write @xmath563 as @xmath564 since each arc @xmath565 lies in @xmath540 , it is contained in a filled box . it follows that @xmath566 ( recall that a @xmath103equivalence between rectangles induces @xmath388relations between all pairs of corresponding points ) . therefore @xmath567 since @xmath390 is the transitive closure of @xmath388 , it follows that @xmath392 . the next proposition provides a bound to how long a flow line can stay in @xmath540 . [ p : arc fibers ] let @xmath253 and let @xmath102 be a small markov partition for @xmath3 . if @xmath402 is not a singular point or a marked point , then @xmath568 is a union of compact arcs and points in @xmath569 . moreover , there exists an integer @xmath570 so that each component of @xmath568 has length strictly less than @xmath571 ( with respect to the flow parameter ) . fix @xmath402 to be a nonsingular , unmarked point . since the set of such points is invariant under @xmath3 , we have that @xmath572 contains no singular or marked points . we need only consider the component @xmath573 containing @xmath385 ( since every component has this form for some @xmath385 ) . recall that lemma [ l : orbit sim ] implies that the @xmath390equivalence class of @xmath385 is precisely @xmath574 for integers @xmath575 . by proposition [ p : prism relation ] , it follows that @xmath563 is equal to the compact arc @xmath576 of length @xmath577 ( if @xmath578 the arc is a point ) . on the other hand , proposition [ p : structure of sim ] implies @xmath579 , where @xmath266 is the unique cell containing @xmath385 in its interior . therefore , the length of the longest arc of intersection of a flow line not passing through a marked point or singular point is @xmath580 which is finite since there are only finitely many cells in @xmath103 . setting @xmath581 completes the proof . in the next proposition , we identify the box @xmath508 with @xmath582 $ ] via the suspension flow as described in section [ s:3-mfd ] . given a space @xmath583 , a subspace @xmath584 and a cell structure @xmath585 on @xmath583 , if @xmath586 happens to be a subcomplex with respect to the cell structure @xmath585 , then we will refer to the induced cell structure on @xmath586 as the _ restriction _ of @xmath585 to @xmath586 , and write it as @xmath587 . [ p : product cells ] for each filled box @xmath508 , the restriction @xmath588 agrees with the product cell structure on @xmath582 $ ] coming from @xmath589 and the cell structure on @xmath590 $ ] that has a single @xmath60cell . in a product cell structure , the cells all have the form @xmath591 , where @xmath522 and @xmath592 are cells of the first and second factors , respectively . thus , the cells of the product @xmath593 $ ] are of the form ( 1 ) @xmath594 $ ] , ( 2 ) @xmath595 , and ( 3 ) @xmath596 . so the proposition is saying that for each cell @xmath266 of @xmath110 in a filled box @xmath508 , there is a cell @xmath522 from @xmath589 so that one of the following holds : @xmath597 so suppose @xmath598 is a filled box . then @xmath599 , and , according to proposition [ p : sim1 and y ] , we have that @xmath600 is cellular with respect to @xmath601 and @xmath602 . since the suspension cells are precisely the suspensions of cells of @xmath123 , which in @xmath197 are exactly the cells of @xmath103 , it follows that all suspension cells are of the form ( 1 ) above . all other cells are surface cells in @xmath197 and @xmath344 . the cells in @xmath197 are of course of the form ( 2 ) . since , again , @xmath603 is cellular with respect to @xmath103 , every cell of @xmath604 is of the form @xmath605 for some cell @xmath522 of @xmath589 , and thus has the form ( 3 ) . let @xmath112 denote @xmath606 and let @xmath607 be the quotient map . according to proposition [ p : prism relation ] , the restriction to @xmath1 of the relation @xmath541 is precisely the relation @xmath390 and so the inclusion of @xmath608 descends to an inclusion @xmath609 , and we have the following commutative diagram : @xmath610^\pi\ar[d ] & t \ar[d ] \\ m \ar[r]^p & n\\}\ ] ] [ p : hat w bounded ] the quotient @xmath112 admits a cell structure @xmath611 so that @xmath37 is cellular with respect to @xmath110 on @xmath50 . moreover , there exists @xmath118 so that @xmath612 has @xmath119bounded complexity . we view @xmath157 as a subset of @xmath112 . first , recall that @xmath540 is the union of the prisms and that @xmath1 is contained in @xmath540 , and observe that all of @xmath540 is mapped ( surjectively ) to @xmath157 . indeed , on any filled box @xmath508 the map to @xmath112 is obtained by first projecting onto @xmath197 using its product structure coming from the flow , then projecting @xmath197 to @xmath157 by @xmath301 . by proposition [ p : product cells ] , this is a cellular map from @xmath613 to @xmath300 , the cell structure on @xmath157 from proposition [ p : w structure ] . let @xmath614 denote the set of unfilled boxes . the @xmath40manifold @xmath50 , with cell structure @xmath110 , is the union of two subcomplexes : @xmath615 note that @xmath540 and @xmath616 are not complementary , as both contain @xmath1 . since both @xmath1 and @xmath540 map onto @xmath157 , it follows that @xmath616 maps onto @xmath112 : @xmath617 we now construct a cell structure @xmath612 on @xmath112 for which @xmath618 is cellular with respect to @xmath619 and for which @xmath157 is a subcomplex with @xmath620 . if we can do this , then since @xmath621 is already cellular , and @xmath50 is the union of the two subcomplexes , it will follow that @xmath37 is cellular . the above description of @xmath112 as the image of @xmath1 and the unfilled boxes indicates the way to build the cell structure @xmath612 on @xmath112 . namely , we start by giving @xmath157 the cell structure @xmath300 and then describe the remaining cells as images of certain cells in @xmath616 . first suppose that @xmath266 is a suspension @xmath60cell that is not contained in @xmath540 . then since any suspension @xmath60cell is contained in the intersection of some set of boxes , it must be that the @xmath60cell meets @xmath540 only in its endpoints . in particular , @xmath37 is injective when restricted to the interior of @xmath266 , and the endpoints map into @xmath622 . we take @xmath623 as a @xmath60cell in @xmath612 . in addition to the @xmath44cells of @xmath300 , we also construct a new @xmath44cell @xmath480 for each suspension @xmath44cell @xmath266 that is not contained in @xmath540 , as follows . the suspension @xmath44cells are contained in intersections of boxes , and so since @xmath266 is not contained in @xmath540 , it meets @xmath540 only in its boundary . therefore , @xmath37 is injective on the interior of @xmath266 . moreover , the boundary contains two suspension @xmath60cells and two arcs that are unions of surface @xmath60cells of @xmath624 . the map @xmath37 sends all the surface @xmath60cells of @xmath625 cellularly into @xmath157 , and is injective on the interior of each cell . each of the two suspension @xmath60cells is either sent to a @xmath60cell constructed in the previous section ( hence injectively on the interior ) , or is collapsed to a @xmath120cell . the cell @xmath480 is essentially @xmath623 , the only difference being the cell structure on @xmath626 has collapsed those suspension @xmath60cells in @xmath625 that are collapsed by @xmath37 in @xmath112 . finally , the @xmath40cells come from the unfilled boxes @xmath614 . each unfilled box has a graph on its boundary coming from the cell structure @xmath110 . moreover , by proposition [ p : hat y ] , this graph has at most @xmath627 vertices , edges , and complementary regions , all of which are disks . these disks are suspension and surface @xmath44cells in @xmath628 . the @xmath40cells in @xmath612 are obtained by first collapsing any suspension @xmath60cells and @xmath44cells in the boundary of the box ( using the product structure from the flow ) that are contained in @xmath540 , then projecting by @xmath37 to @xmath112 . this collapsing produces a new graph in the boundary of the @xmath40cell , but the new graph has no more edges , vertices or complementary regions than the original one . moreover , all complementary components are still disks ; see figure [ f:3cellcollapse ] for a picture of the @xmath40cell and the new cell structure on the boundary . with this new cell structure on the boundary , the restriction of @xmath37 to the interior of each cell is injective . we also observe that each of these @xmath40cells has @xmath518bounded complexity . it is straightforward to check that this is the desired cell structure ; @xmath37 is cellular on @xmath616 , and by construction @xmath157 is a subcomplex with @xmath620 . therefore , @xmath37 is cellular on all of @xmath629 as discussed above , and we have proven the first part of the proposition . now we find a @xmath630 so that @xmath612 has @xmath119bounded complexity . first , the number of @xmath40cells is precisely the number of unfilled boxes . each unfilled box is given by @xmath631 where @xmath106 is a terminal rectangle in its @xmath103equivalence class . therefore , the unfilled boxes correspond precisely to the @xmath103equivalence classes of rectangles . by corollary [ c : relation 1 prop ] , there are at most @xmath632 unfilled boxes , and so at most @xmath373 @xmath40cells . as mentioned above , each of these @xmath40cells has @xmath518bounded complexity . the cell structure @xmath300 has @xmath473bounded complexity by proposition [ p : w structure ] , so there are at most @xmath474 @xmath60cells in @xmath300 . each @xmath60cell of @xmath612 not in @xmath300 is contained in the boundary of some @xmath40cell . since these have @xmath518bounded complexity , and since there are at most @xmath373 of these , it follows that there are at most @xmath633 @xmath60cells in @xmath612 that are not in @xmath300 . therefore , there are at most @xmath634 @xmath60cells in @xmath612 . similarly , the number of @xmath44cells is bounded by the number of @xmath44cells in @xmath300 plus the number of @xmath44cells in each of the unfilled boxes . a count as in the previous paragraph implies that there are at most @xmath635 @xmath44cells in @xmath612 . finally , the @xmath44cells of @xmath300 have @xmath474bounded complexity by proposition [ p : w structure ] , and each of the suspension @xmath44cells has @xmath518bounded complexity since it is in the boundary of a @xmath40cell , which has @xmath518bounded complexity . therefore , setting @xmath636 it follows that @xmath612 has @xmath119bounded complexity . let @xmath637 the following is immediate from propositions [ p : bounded complexity finite ] and [ p : hat w bounded ] . [ c : compact finiteness ] the set @xmath117 is finite . we continue with the notation from the previous section : @xmath638 , @xmath540 is the union of @xmath1 and the prisms , and @xmath607 is the quotient defined by collapsing arcs of flow lines in @xmath540 to points . let @xmath114 denote @xmath639 , the complement in @xmath112 of the singular and marked subcomplexes . we also write @xmath640 for the restriction . the main goal of this section is to prove the following . recall that a 3manifold is _ irreducible _ if every 2sphere in the 3manifold bounds a 3ball . we will deduce theorem [ t : quotient htpc homeo ] from waldhausen s theorem ( the version we will use can be found in ( * ? ? ? * corollary 13.7 ) ) : [ t : wald ] suppose @xmath641 and @xmath642 are compact , orientable irreducible 3manifolds with nonempty boundary . if @xmath643 is a map such that @xmath644 is an isomorphism and @xmath645 is an injection for each component @xmath646 and the component @xmath647 containing @xmath648 , then @xmath649 is homotopic to a homeomorphism . by corollary [ c : compact finiteness ] , @xmath117 is a finite set of cw homeomorphism types of cell complexes . therefore , since each @xmath114 is obtained by removing a @xmath60subcomplex from a compact @xmath40complex @xmath650 , it follows that the set @xmath651 is finite . but by theorem [ t : quotient htpc homeo ] , this set is equal to @xmath16 . let @xmath652 denote the fibration coming from the mapping torus description of @xmath67 . let @xmath653 be a volume form so that @xmath654 , and let @xmath655 be the pullback ( which on @xmath656 $ ] we can take to be @xmath657 ) . this @xmath60form @xmath658 represents the poincar dual of the class represented by @xmath659 . integration of @xmath658 defines an epimorphism @xmath660 , and we let @xmath661 denote the infinite cyclic cover of @xmath67 determined by @xmath662 . alternatively , @xmath166 is simply the cover corresponding to @xmath663 , and so is obtained by unwrapping the bundle . therefore , we have @xmath664 we make this diffeomorphism explicit as follows . first , choose a lift @xmath665 of the embedding @xmath659 . the flow @xmath108 lifts to a flow @xmath666 on @xmath667 . we now define a map @xmath668 by @xmath669 we use this diffeomorphism to identify @xmath667 with @xmath670 . the covering group for @xmath671 is infinite cyclic . let @xmath672 be the generator that induces a translation by @xmath673 in the second factor of @xmath674 . with respect to the product structure , we have @xmath675 writing the last formula in terms of the flow @xmath666 , we obtain @xmath676 and iterating , we obtain @xmath677 we will study the quotient @xmath114 of @xmath67 by looking at the associated quotient of @xmath667 . let @xmath678 , and let @xmath679 . for any @xmath680 , we consider the flow line @xmath681 since @xmath682 is a lift of the flow line @xmath569 , proposition [ p : arc fibers ] implies that @xmath683 is a union of compact arcs ( and points ) of length less than @xmath570 ( measured with respect to the flow parameter ) . for @xmath684 define @xmath685 if there exists @xmath680 so that @xmath686 are in the same component of @xmath683 . set @xmath687 and let @xmath688 be the quotient map . we will now give an alternative description of @xmath689 as an open subset of @xmath670 ; we will realize it as the image of a smooth map @xmath690 . let @xmath691 be a smooth function that is positive on @xmath692 and is identically zero on @xmath693 . let @xmath694 , and @xmath695 . since @xmath696 is a pullback , it is invariant by @xmath672 : @xmath697 let @xmath657 denote the 1form on @xmath674 obtained by pulling back the volume form on @xmath25 , and observe that on @xmath670 we have @xmath698 now define @xmath699 by @xmath700 where the notation in the integral means the integral of @xmath696 over the path @xmath701 as @xmath702 runs from @xmath120 to @xmath703 . since @xmath696 is smooth , the map @xmath649 is smooth . set @xmath704 we remark that @xmath583 may or may not be all of @xmath705 . first observe that if @xmath708 , then @xmath709 since @xmath710 . it follows that we can only have @xmath711 if @xmath464 and @xmath549 both lie on @xmath682 for some @xmath385 . therefore , it suffices to show that for any @xmath680 and for all @xmath712 , we have @xmath713 if and only if @xmath685 . to prove this , we observe that for any flow line @xmath682 , the restriction @xmath714 of @xmath696 to @xmath682 is zero precisely on the intersection @xmath683 . therefore , @xmath715 is constant on each component of @xmath716 and monotone ( increasing ) on the complementary intervals . it follows that the fibers of @xmath715 are precisely the @xmath706equivalence classes . [ p : covering v ] the image @xmath721 is an open subset of @xmath670 , and is homeomorphic to @xmath670 . moreover , @xmath722 acts properly discontinuously and freely on @xmath583 , and with respect to this product structure , @xmath719 is given by @xmath105 on the first factor . let @xmath570 be the constant from proposition [ p : arc fibers ] . as mentioned above ( immediately after the definition of @xmath682 ) , the integer @xmath571 is greater than the length of the component of any flow line @xmath723 . in particular , it follows that @xmath724 for all @xmath402 . the region between @xmath69 and @xmath729 in @xmath705 is homeomorphic to a product . indeed , it is exactly the region between the graphs of the zero function on @xmath69 and the smooth positive function @xmath730 we denote this region by @xmath131 . observe that @xmath731)\ ] ] and so @xmath732 . the region @xmath131 is a fundamental domain for the action of @xmath733 on @xmath583 , and @xmath583 is equal to @xmath734 with @xmath733 acting properly discontinuously . since @xmath733 is contained in @xmath735 with finite index , it follows that @xmath735 also acts properly discontinuously . because @xmath735 is torsion free , its action on @xmath583 is also free . any homeomorphism from @xmath131 to @xmath736 $ ] that is the identity on the surface factor uniquely extends to a homeomorphism @xmath737 that conjugates the action of @xmath733 to the action of @xmath738 . as the homeomorphism is the identity on the first factor , the conjugate of @xmath719 ( which is different from @xmath672 ) restricts to @xmath105 on the first factor . the map @xmath739 is a proper continuous surjection , and hence is a quotient map . also , by definition , @xmath740 is a quotient map . since the fibers of these two maps are the same ( lemma [ l : same fibers ] ) , it follows that the quotients are homeomorphic . let @xmath741 be the homeomorphism for which @xmath742 . we use the homeomorphism @xmath743 and the homeomorphism @xmath744 from proposition [ p : covering v ] to identify @xmath689 with @xmath670 . conjugating by this homeomorphism we obtain an action of @xmath735 on @xmath689 . we abuse notation and simply refer to the conjugate homeomorphism by the same name : @xmath745 now , @xmath114 is the image of @xmath67 via the quotient map @xmath640 , and since the covering map @xmath746 is also a quotient map , the composition @xmath747 is a quotient map . we also have the covering map @xmath748 which is a quotient map , and hence so is @xmath749 . since the fibers of @xmath750 are the same as those of @xmath749 , it follows that @xmath751 general covering space theory implies that @xmath752 fits into a short exact sequence : @xmath753 the monodromy @xmath754 is given by @xmath755 . on the other hand , the isomorphism @xmath756 induced by projecting onto the first factor , conjugates @xmath757 to @xmath758 by proposition [ p : covering v ] . therefore , @xmath752 is an extension of @xmath759 by @xmath663 with monodromy @xmath760 given by @xmath761 . the fundamental group @xmath762 also has such a description . since @xmath759 is free , each sequence splits , and it follows that @xmath762 is isomorphic to @xmath752 . moreover , the map @xmath640 induces this isomorphism @xmath763 as can be seen in the lift @xmath764 . since @xmath114 is covered by @xmath670 and @xmath69 is a hyperbolic surface , it follows that the universal cover of @xmath114 is @xmath765 . therefore @xmath114 is irreducible . furthermore , @xmath114 is the complement of a @xmath60subcomplex of the cell complex @xmath112 , which has @xmath119bounded complexity by proposition [ p : hat w bounded ] . since @xmath112 can be subdivided into a simplicial complex ( see the proof of proposition [ p : bounded complexity finite ] ) , it follows that @xmath114 is tame , that is , it is homeomorphic to the interior of a compact 3manifold with boundary . because @xmath114 is tame , we can remove a product neighborhood of the ends to produce a _ compact core _ @xmath766 for @xmath114 , which is a compact submanifold for which the inclusion is a homotopy equivalence . since @xmath37 is a proper map , @xmath767 is a compact subset of @xmath67 . on the other hand , @xmath67 is also tame , and so we can also remove product neighborhoods of the ends of @xmath67 to arrive at a compact core @xmath768 for @xmath67 that contains @xmath767 . using the product structure on the complement of @xmath768 in @xmath67 and of @xmath766 in @xmath114 , there are strong deformation retractions @xmath769 and @xmath770 . the map @xmath771 is therefore homotopic to @xmath37 , and moreover the restriction @xmath772 , satisfies the hypotheses of waldhausen s theorem ( theorem [ t : wald ] ) : the only thing to verify is that the boundary subgroups are mapped injectively , but that is clear from the construction of @xmath37 . therefore , @xmath773 is homotopic to a homeomorphism . using this homotopy and the product neighborhoods of the ends one can construct a homotopy from @xmath774 , and hence from @xmath37 , to a homeomorphism . ulrich oertel . homology branched surfaces : thurston s norm on @xmath775 . in _ low - dimensional topology and kleinian groups ( coventry / durham , 1984 ) _ , volume 112 of _ london math . lecture note ser . _ , pages 253272 . cambridge univ . press , cambridge , 1986 .	the main result of this paper is a universal finiteness theorem for the set of all small dilatation pseudo - anosov homeomorphisms @xmath0 , ranging over all surfaces @xmath1 . more precisely , we consider pseudo - anosovs @xmath0 with @xmath2 bounded above by some constant , and we prove that , after puncturing the surfaces at the singular points of the stable foliations , the resulting set of mapping tori is finite . said differently , there is a finite set of fibered hyperbolic 3manifolds so that all small dilatation pseudo - anosovs occur as the monodromy of a dehn filling on one of the 3manifolds in the finite list , where the filling is on the boundary slope of a fiber .
the rotation induced oblateness of astronomical bodies is a classical problem in newtonian and celestial mechanics ( for the early history , see todhunter @xcite ) . it has twice played an important role in the history of science . in the early eighteenth century measurements indicated a prolate shape of the earth , in strong conflict with the newtonian prediction . this was later shown to be wrong by more careful measurements by maupertuis , clairaut , and celsius in northern sweden in 1736 . then , in 1967 , measurements of the solar oblateness were published and according to these it was much larger than the sun s surface angular velocity would explain . the confirmation of general relativity by mercury s perihelion precession would then be lost . also this problem is now gone and the modern consensus is that the solar oblateness is too small to affect this classic test of general relativity @xcite . the subject of the flattening of rotating astronomical bodies is thus quite mature . the classical theory is due mainly to clairaut , laplace , and lyapunov . also radau , darwin , de sitter and many others have made important contributions . more recent accounts of the theory can be found in , for example , jeffreys @xcite , zharkov et al.@xcite , cook @xcite , moritz @xcite and , partly , in chandrasekhar @xcite . some pedagogical efforts can be found in murray and dermott @xcite , or in kaula @xcite . as is plain from these references the theory is quite involved . only the unrealistic assumption that the body is homogeneous gives compact analytical results . otherwise a specified radial density distribution is needed and one resorts to cumbersome series ( multipole ) expansions , or purely numerical methods , for quantitative results . here we will present analytical results based on the assumption that the body consists of a central point mass surrounded by a homogeneous fluid , the so called point core model . by varying the relative mass of the fluid and the central point particle one can interpolate between the extreme limits of newton s homogeneous body and the roche model @xcite with a dominating small heavy center . the point core model goes back to the work of g. h. darwin . more recently it has been used to study the shape of outer planet moons , see hubbard and anderson @xcite , dermott and thomas @xcite . apart from the basic point core approximation ( i ) several further approximations are assumed here . these are : ( ii ) that the shape is determined by hydrostatic equilibrium and ( iii ) that the shape is ellipsoidal . these are not consistent . according to hamy s theorem , see moritz @xcite , the exact shape is not ellipsoidal , so we regard an ellipsoidal shape as a constraint and find the equilibrium shape , among these , by minimizing the energy . a further approximation ( iv ) is the neglect of differential rotation . on the other hand fixed volume ( `` bulk incompressibility '' ) need not be assumed ; the equilibrium volume problem separates from the shape problem . in spite of these approximations , which are standard in the literature , the mathematics can be quite involved . in this article i hope to clarify and simplify it as much as possible . mathematically our model then becomes a three degree of freedom mechanical system for which we can calculate the kinetic and potential ( gravitational plus centrifugal ) energies exactly . multipole expansions in terms of spherical harmonics are not needed . the three degrees of freedom correspond to the three semi - axes of the ellipsoid ( @xmath3 ) but these are transformed to three generalized coordinates that describe size ( or volume ) , @xmath4 , spheroidal , @xmath5 , and triaxial , @xmath6 , shape changes , see eq . ( [ eq.def.xi.eta ] ) . the statics problem of equilibrium shape is solved by minimizing the potential energy , @xmath7 for a fixed @xmath4 , eq.([eq.energy.xi.tau ] ) . for slow rotation the shape will not be triaxial so @xmath8 . finding the shape , or flattening , is then only a matter of minimizing a dimensionless potential energy , @xmath9 given by the last two terms of eq . ( [ eq.energy.xi ] ) . here @xmath10 is defined by eq . ( [ eq.basic.phi.xi ] ) and plotted in fig . [ fig1 ] while @xmath11 is a constant that depends on rotational parameter @xmath2 and dimensionless moment of inertia @xmath12 , eq.([eq.kappa2.nu ] ) . the root @xmath13 of the equation @xmath14 thus gives the flattening . this root is to very high accuracy given by @xmath15^{-1/6 } , \ ] ] see eqs . ( [ eq.maclaurin.approx.inv ] ) and ( [ eq.eccentricity.relation.xi ] ) , for the rotational parameters that can be found in the solar system . from this result , simple formulas , eqs.([eq.maclaurin.approx.relation.q ] ) ( [ eq.maclaurin.approx.relation.epsilon ] ) , relating the observables , rotation parameter , @xmath2 , gravitational quadrupole , @xmath1 , and excentricity squared , @xmath16 , are obtained . these appear to be , partly , new , and their usefulness is demonstrated by comparing with empirical data for the sun and the rotating planets of the solar system . variational methods have been used before to study similar problems , see for example abad et al . denis et al.@xcite have pointed out that variational methods generally fail to provide estimates of their accuracy . therefore the agreement of our formulas with empirical data , as demonstrated in table /cite(table ) , is important and demonstrates that our model catches the essential physics of rotational flattening . finally small amplitude oscillations near the equilibrium are investigated , starting from the lagrangian @xmath17 for our three degree of freedom model system . this gives useful insight into the physics of free stellar or planetary oscillations and their coupling to rotation . our approximations are , however , too severe for these results to be of quantitative interest . assume that @xmath18 are rectangular cartesian coordinates in three - dimensional space of a point with position vector @xmath19 an ellipsoid , with semi - axes @xmath3 , is the solid defined by @xmath20^{1/2 } \le 1 .\ ] ] if we put @xmath21 and define @xmath22^{1/2}\ ] ] the ellipsoid is the set of points that fulfill @xmath23 where @xmath4 is the geometric mean ( or volumetric ) radius @xmath24{abc}$ ] . the formula for the volume of an ellipsoid now gives @xmath25 and we see that changes of @xmath5 and @xmath6 do not change the volume , only the shape of the ellipsoid . in what follows we will also consider the special case of spheroids . a spheroid is an ellipsoid with two equal semi - axes . we take @xmath26 . this means that we take @xmath8 in the formulas above so that @xmath27 and ( [ eq.ellipsoid.e.def ] ) becomes @xmath28^{1/2}.\ ] ] a spheroid is thus defined as the set of points @xmath29 that fulfill @xmath30 here @xmath4 is the ( geometric ) mean radius and @xmath5 is a parameter that determines the shape in such a way that @xmath31 corresponds to a prolate spheroid , @xmath32 to a sphere , and @xmath33 to an oblate , or flattened , spheroid . more familiar parameters used to define the shape of a spheroid are the ellipticity @xmath0 and the eccentricity @xmath34 . the ellipticity is defined by @xmath35 this is sometimes also called the ( geometric ) oblateness or the flattening ( denoted @xmath36 ) . solving this equation , and expanding around @xmath37 , we have @xmath38{1-\epsilon}}=1+\frac{1}{3}\epsilon + \frac{2}{9}\epsilon^2+o(\epsilon^3).\ ] ] one notes that @xmath0 is positive for oblate and negative for prolate spheroids respectively . the eccentricity , @xmath39 , is defined as the usual eccentricity of the ellipse that is the intersection of the spheroid and a plane containing the z - axis . one finds that @xmath40 gives the eccentricity of these ellipses in the oblate case . for the prolate case @xmath41 consider a non - rotating spherically symmetric body consisting of point particles with masses @xmath42 and position vectors @xmath43 assume that rotation induces a deviation of the positions so that the new positions , @xmath44 , are given by @xmath45 this implies that we assume that the elastic displacement field can be parameterized by the shape parameters @xmath46 with non - rotating positions corresponding to @xmath47 . one notes that points @xmath48 that obeyed @xmath49 i.e. were found on a sphere of radius @xmath50 , move to the surface given by @xmath51 the sphere is thus assumed to deform to an ellipsoid . assume that a body initially is spherically symmetric and has a mass density @xmath52 , where @xmath53 is the ordinary distance from the center . it is tempting to assume that a natural deformation of the body when it starts to rotate is to a spheroidal shape in such a way that the originally spherical equidensity surfaces , @xmath54 , deform to similar spheroidal surfaces given by @xmath55 . unfortunately this is an approximation for real bodies . for a density that increases towards the center , the ellipticity of the equidensity surfaces decrease with @xmath53 , in a manner described by clairaut s equation , see @xcite . note , however , that for a constant density , or a constant density with a central point mass , the assumption is not approximate , since any surface is an equidensity surface , only the shape of the outer surface matters . thus we now assume such a mass distribution . to be precise we assume that a mass @xmath56 is distributed with constant density @xmath57 inside the surface , @xmath58 , and that the point mass @xmath59 is at the center , @xmath60 , of the ellipsoid . the density @xmath61 is thus given by @xmath62 where @xmath63 is the three - dimensional dirac delta function . if we put @xmath42 for the masses of the particles , the inertia tensor of the body is given by @xmath64 where we have put , @xmath65 , and @xmath66 is the usual kronecker delta . the inertia tensor for the non - rotating body is diagonal with all diagonal elements equal and given by @xmath67.\ ] ] here we have introduced the total mass , @xmath68 as well as the dimensionless radius of gyration @xmath69 . the square of this parameter , @xmath12 , also appears in the literature denoted @xmath70 ( or @xmath71 , or simply @xmath11 ) and is also called the moment of inertia coefficient . we calculate the dimensionless moment of inertia @xmath12 for the undeformed @xmath61 of eq . ( [ eq.mass.distribution ] ) . using eq . ( [ eq.def.mom.of.inert ] ) and @xmath72 we find @xmath73 where @xmath74 and eq . ( [ eq.rho.h ] ) have been used . this is the usual moment of inertia for a solid sphere of mass @xmath75 . this implies that @xmath76 is the squared ( dimensionless ) radius of gyration of the undeformed body . note that @xmath77 since the model can go between the limits of a point ( @xmath78 ) and a homogeneous sphere ( @xmath79 ) . for the deformed body one gets , using ( [ eq.position.vectors.dep ] ) , @xmath80 @xmath81 is often denoted by @xmath82 . for spheroids ( @xmath8 ) the other moments of inertia become @xmath83 for these one frequently finds the notation @xmath84 , and @xmath85 , respectively , in the literature . the quadrupole tensor , @xmath86 , is given in terms of the inertia tensor by @xmath87 it is identically zero for the undeformed body and diagonal for the same axes as the inertia tensor . for the spheroidal ( @xmath8 ) body its components are given by @xmath88 and @xmath89 thus @xmath90 ( or @xmath91 is positive for oblate and negative for prolate shapes . one measure of the oblate shape that is of great interest is the , so called , gravitational quadrupole @xmath92 . this dimensionless quantity , gives the deviation of the gravitational potential , @xmath93 , outside the body , from the simple @xmath94-dependence of spherically symmetric bodies . the first two terms in the multipole expansion of this potential are , see e.g. stacey @xcite , @xmath95,\ ] ] where @xmath96 is the polar angle ( colatitude ) in spherical coordinates , @xmath97 the total mass , and @xmath98 the equatorial radius of the body . one can show @xcite that @xmath92 of formula ( [ eq.grav.pot.j2 ] ) is given by @xmath99 . use of @xmath100 given in ( [ eq.quadrupole.comp.xx ] ) , and @xmath101 , then gives @xmath102 for the quadrupole . here we have also used @xmath101 and eq.([eq.eccentricity.relation.xi ] ) . note that @xmath103 is independent of @xmath4 , but directly proportional to the eccentricity squared . assuming the oblateness to be small , and due to hydrostatic equilibrium in the combined gravitational and centrifugal force fields , one can derive a formula connecting @xmath92 with the angular velocity and the ellipticity . one uses that the shape of the surface is determined by the hydrostatic equilibrium equation @xmath104 where @xmath105 is pressure , @xmath61 mass density , and @xmath106 is the potential of the centrifugal force in a system rotating with angular velocity @xmath107 about the z - axis . this means that the surface of the body must be an equipotential surface of @xmath108 . the constancy of @xmath109 on the surface of the body can be used to derive the approximate relationship @xmath110 @xcite . here we have introduced the ratio @xmath2 of equatorial centrifugal to gravitational acceleration , @xmath111 here @xmath112 is the mean density . instead of @xmath2 the notation @xmath113 occurs in the literature , see zharkov et al.@xcite , who reserve @xmath2 for the corresponding quantity with the equatorial radius @xmath98 replacing the mean radius @xmath4 ( recall @xmath114 ) . the advantage of the definition ( [ eq.acc.centrifug.grav.ratio ] ) is that it makes @xmath2 independent of flattening . the relationship ( [ eq.j2.alpha.epsilon ] ) , is only valid to first order in small quantities , so here the difference does not matter , but we will go to higher order below , so take notice . this relationship between the three quantities @xmath2 ( or @xmath113 ) , @xmath0 ( or @xmath36 ) , and @xmath1 , can thus be used to test the hypothesis of hydrostatic equilibrium empirically . this will be done below , using more exact results . for jupiter and saturn these will improve significantly on ( [ eq.j2.alpha.epsilon ] ) . to study a rotating body , we go to the rotating system , and assume that all particles are affected by the centrifugal and gravitational potential energies . if the system is rotating rigidly about the z - axis with angular velocity vector @xmath115 the potential energy of particle @xmath116 in the centrifugal force field is @xmath117 . the total centrifugal potential energy of the body ( system of particles ) is @xmath118 this is the negative of the rotational kinetic energy . using ( [ eq.mom.of.inert.kappa ] ) we can write @xmath119 for the total potential energy of the particles in the centrifugal potential . the gravitational potential from an ellipsoid ( [ eq.ellipsoid.ec.def ] ) , of constant density @xmath120 , at an interior point is @xcite @xmath121 where @xmath122 . since @xmath123 is the interaction energy of the homogeneous ellipsoid with a point particle of mass @xmath59 at the origin one finds the expression @xmath124 for this interaction energy . here @xmath125 and eq . ( [ eq.rho.h ] ) has been used . an elementary calculation based on ( [ eq.grav.pot.of.hom.ellipsoid.interior ] ) , see landau and lifshitz @xcite , also shows that the gravitational self energy of the ellipsoid is @xmath126 use of the definitions ( [ eq.def.xi.eta ] ) and the substitution @xmath127 gives @xmath128[(\xi-\tau)^2+u][(\xi^{2}-\tau^{2})^{-2}+u]}}.\ ] ] if we define @xmath129[(\xi-\tau)^2+s][(\xi^{2}-\tau^{2})^{-2}+s]}},\ ] ] we find that the total gravitational energy of the mass distribution ( [ eq.mass.distribution ] ) is given by @xmath130 where @xmath131 . we define the constant @xmath132 , the dimensionless gravitational radius , through @xmath133 comparing with ( [ eq.tot.grav.energ ] ) and using @xmath134 together with the result ( [ eq.kappa2.of.ms ] ) , a small calculation reveals that the dimensionless gravitational radius , @xmath132 , is given by @xmath135 it is interesting to note that the gravitational energy of a density ( [ eq.mass.distribution ] ) , with fixed mass @xmath97 and radius @xmath4 , is minimized for @xmath136 ( this is incidentally very close earth s value ) . using the definition ( [ eq.grav.energy.triaxial.chi ] ) it is easy to find the taylor expansion of @xmath137 around @xmath47 . this gives @xmath138 so the sphere really corresponds to a minimum in the gravitational energy . higher order terms are easily generated by computer algebra programs , but we will not need them here . to study the behavior further away from the minimum one might use , instead of ( [ eq.grav.energy.triaxial.chi ] ) , an excellent pad approximation of the ellipsoidal energy derived by dankova and rosensteel @xcite . near the spherical minimum it is almost indistinguishable from the exact expression . for spheroids we define @xmath139 to be the dimensionless gravitational potential energy . for @xmath8 the integral in ( [ eq.grav.energy.triaxial.chi ] ) can be found in terms of elementary functions . several different alternative expressions exist and they are always different depending on whether @xmath5 is greater or less than unity . however , noting that , @xmath140 , the two expressions in @xmath141 are equivalent and can be used for all @xmath142 . these are really the same real function if only the correct branch is chosen when passing through @xmath32 . series expansion , of either alternative , around @xmath32 gives @xmath143.\ ] ] this function , which is shown in figure [ fig1 ] , has the following properties . the prolate limit , @xmath144 , of an infinite line is zero : @xmath145 . the oblate limit , @xmath146 , an infinite circular disc , is also zero : @xmath147 . the sphere ( @xmath32 ) corresponds to a minimum of @xmath10 so that @xmath148 , @xmath149 , and @xmath150 . the total energy , @xmath151 , of a static body ( star or planet ) will , apart from the centrifugal energy , @xmath152 of eq . ( [ eq.centrifug.energy.1 ] ) , and gravitational energy , @xmath153 of eq . ( [ eq.grav.energy.triaxial.2 ] ) , also consist of some volume dependent energy , @xmath154 . we can then write this total energy it in the form @xmath155 by minimizing this energy we can find out whether a system tends to be triaxial or spheroidal . for small angular velocities the spheroidal case is known to be relevant @xcite . we thus put @xmath8 here and get @xmath156 using this we can find out if the body will be prolate or oblate and how this is affected by rotation . since we concentrate on stars and planets surface forces and other neglected shape dependent forces are normally very small compared to the gravitational and centrifugal energies . the equilibrium equations can be written @xmath157 @xmath158 where we have denoted differentiation by @xmath5 with a prime . if we multiply the first equation by @xmath4 and the second by @xmath5 and subtract we get , after rearrangement , @xmath159.\ ] ] this is essentially an equation of hydrostatic equilibrium . note that it determines the equilibrium @xmath4-value with no direct reference to the angular velocity @xmath107 . the second equation , ( [ eq.de.dxi ] ) , can be written @xmath160 and this expresses a balance of the shape dependent part of the gravitational force with the centrifugal force . if we put @xmath161 this turns into the concise expression @xmath162 since @xmath149 this will cause a shift of the shape from the spherical equilibrium , at @xmath32 , to @xmath33 , as long as @xmath163 . use of ( [ eq.tot.grav.energ.nu.ch ] ) shows that @xmath164 the equilibrium @xmath5 is thus entirely determined by @xmath2 and the moment of inertia @xmath12 . to investigate eq . ( [ eq.force.bal.grav.surf.centrifug.2 ] ) we use the second variant for @xmath10 in eq . ( [ eq.basic.phi.xi ] ) and get , after some algebra , @xmath165 .\ ] ] this looks simpler if we introduce the new variable @xmath166 . use of eq . ( [ eq.eccentricity.relation.xi ] ) shows that @xmath167 so that @xmath168 can also be regarded as a function of the eccentricity , @xmath169 . now we get @xmath170/\eta^3 .\ ] ] for the homogeneous case @xmath171 and this formula , written in terms of the eccentricity by means of the identity , @xmath172 , is equivalent to the celebrated maclaurin s formula @xcite from 1742 . taylor expansion of the maclaurin function around @xmath173 only contains positive even powers of @xmath34 . the coefficient of @xmath174 vanishes so a very good approximation for this function is implied by @xmath175/\eta^3 = \frac{4}{15 } e^2 \left ( 1+\frac{1}{7}e^2 \right ) + o(e^8).\ ] ] the two term approximation is nearly perfect for all planetary @xmath34-values , see figure [ fig2 ] . saturn has maximum @xmath176 and the absolute error at this value is @xmath177 and the relative error is @xmath178 . the relative error even remains below @xmath179 up to @xmath180 . using this approximation we must solve @xmath181 and solving this for @xmath16 gives @xmath182 use of eq . ( [ eq.j2.expression ] ) gives @xmath183 . insertion of this into ( [ eq.kappa2.nu ] ) eliminates @xmath12 from @xmath184 . some algebra then leads to the equation @xmath185 for @xmath2 . this equation can be solved for @xmath1 or for @xmath16 to give @xmath186,\end{aligned}\ ] ] respectively . these results , together with @xmath187 are thus more exact versions of eq . ( [ eq.j2.alpha.epsilon ] ) . already the approximation @xmath188 , to the maclaurin function , gives only a @xmath189 error for saturn , and simpler formulae , but use of these more exact versions removes all numerical considerations from the problem . table [ table ] compares modern data with the predictions of eqs.([eq.maclaurin.approx.relation.q ] ) ( [ eq.eps.e2 ] ) . observed data for @xmath2 , @xmath0 , and @xmath1 from lodders and fegley @xcite are given in the first ( horizontal ) row of data for each of the major rotating planets . in the same line @xmath16 calculated from @xmath190 is then given . the last value in the row is the dimensionless moment of inertia @xmath191 as given by lodders and fegley @xcite . these values are based on calculation . for original sources the reader may consult the planetary scientist s companion by lodders and fegley @xcite . for each pair of the values @xmath2 , @xmath1 , and @xmath16 , in the first row , the third value is calculated using ( [ eq.maclaurin.approx.relation.q ] ) ( [ eq.maclaurin.approx.relation.epsilon ] ) and the results are displayed in the second row of data for each planet in table [ table ] . @xmath0 is then found from equation ( [ eq.eps.e2 ] ) . for jupiter and saturn there is a third row of data in table [ table ] . this row is analogous to the second row except that the data are calculated from observational pairs of data using the first order formula ( [ eq.j2.alpha.epsilon ] ) so that @xmath2 is given by observational @xmath0 and @xmath1 according to @xmath192 , @xmath0 is found directly from ( [ eq.j2.alpha.epsilon ] ) using the observational @xmath2 and @xmath1 and so on . these rows illustrate the fact that the non - linear formulas of this work are more accurate than first order results , in spite of the point core approximation . even for the earth the first order formula gives discrepancies in the third digit . table [ table ] shows clearly that the observational data and theory are in reasonable agreement for the planets with the exception of mars and uranus . the perfect agreement for the earth is probably due to the fact that the oblateness is determined from the shape of the geoid ( the equipotential surface at sea - level ) . for mars the disagreement may depend on the fact that oblateness is determined from geometrical shape rather than from the shape of an equipotential surface also , of course , the assumption of hydrostatic equilibrium may not be perfect for mars . the problems with the uranus data , and to some extent those for neptune , are probably due to observational uncertainty . instead of eliminating the moment of inertia ( or mass ratio ) @xmath12 of eq . ( [ eq.kappa2.of.ms ] ) from the equations as done above one can retain it and try to use observed data to get information about @xmath12 . series expansion of eq.([eq.maclaurin ] ) with @xmath168 expressed in terms of appropriate variables gives @xmath193 for the flattening , or oblateness , expressed in terms of @xmath2 and the dimensionless moment of inertia . for a homogeneous body @xmath194 so @xmath195 and the old newtonian result , @xmath196 , is obtained , to first order . in the opposite ( roche ) limit of dominating central mass ( @xmath197 ) and one finds @xmath198 . in a similar way we get from @xmath92 of eq . ( [ eq.j2.expression ] ) that @xmath199 is the dimensionless quadrupole moment of the rotating body . for the homogeneous ( @xmath194 ) body this implies @xmath200 . in the opposite limit of @xmath201 one finds that @xmath92 goes to zero . our results ( [ eq.ellipticity.of.kappa2 ] ) and ( [ eq.j2.expression.3 ] ) , to first order , are @xmath202 and imply that @xmath203 . these may be compared to similar expressions derived by g. h. darwin using the radau equation @xcite @xmath204 these two have the same values at @xmath194 , the homogeneous sphere , and the same derivatives with respect to @xmath12 at this point , as ( [ eq.ellipt.kappa2 ] ) and ( [ eq.j2.kappa2 ] ) , respectively . the simple expressions ( [ eq.ellipt.kappa2 ] ) and ( [ eq.j2.kappa2 ] ) , derived here , however , also have natural limits for @xmath197 , in contrast to e.g. ( [ eq.j2.kappa2.darw ] ) , which becomes negative for @xmath205 . we now use eq . ( [ eq.j2.expression ] ) i.e. @xmath183 and eq . ( [ eq.maclaurin.approx.relation.epsilon ] ) for @xmath206 to calculate @xmath12 in terms of the observed @xmath2 and @xmath1 . the last entry in the second row for each planet of table [ table ] gives @xmath12 , @xmath207 as calculated from observational , first row , @xmath2 and @xmath1 . since @xmath2 and @xmath1 are easier to measure accurately than flattening this should give more reliable values than using the observed ellipticity directly . in this way one obtains information about the radial mass distribution in the interior of the planet from purely external data . the resulting @xmath12-values are given last in the second line for each planet in table [ table ] . the data obtained in this way are compared with published and tabulated @xmath12 values for the planets ( lodders and fegley @xcite ) . for earth and mars the agreement is reassuring . for the outer planets there is naturally a lot of uncertainty but at least the increasing trend from a minimum at saturn is a common feature . one problem with the outer planets , and with the sun in particular , is to decide what angular velocity , @xmath107 , should be used to find @xmath2 . since these bodies have differential rotation some average must be used . there is a well defined theoretical average angular velocity ( essn @xcite ) but it is not directly observable . once an average angular velocity has been selected for the sun one can use the present theory and a theoretical @xmath12-value ( @xmath208 ) to estimate the solar oblateness and @xmath1 . the definition of average angular velocity in @xcite indicates that an angular velocity near the radius of gyration @xmath209 should be appropriate . the further assumption that angular velocity is constant on coaxial cylinder surfaces ( kippenhahn and weigert @xcite , ulrich and hawkins @xcite ) indicates that an angular velocity near the poles of the sun rather than equatorial angular velocity is relevant . a handbook value @xcite for the polar angular velocity of the sun is @xmath210rad s@xmath211 . use of this gives the @xmath2-value ( @xmath212 ) in table [ table ] . the precise numbers are not important here . what is important is that , because of the small @xmath12 , an angular velocity near the pole rather than an equatorial angular velocity is used . otherwise the flattening will be exaggerated . use of @xmath213 and @xmath214 from lodders and fegley @xcite , together with formulas ( [ eq.ellipt.kappa2 ] ) and ( [ eq.j2.kappa2 ] ) , gives the results in the top row of table [ table ] , that is , @xmath215 , and @xmath216 . these values are , at least , of the same order of magnitude as the currently best motivated values @xcite which are roughly @xmath217 and @xmath218 . we now wish to study the small amplitude motions of a gravitating rotating nearly spherical body . we wish to know how the radius @xmath4 is affected by the rotation so we take @xmath4 to denote the constant non - rotating equilibrium radius and let @xmath219 be the variable radius . @xmath220 is thus a dimensionless variable which is equal to unity at the undeformed geometry . we parameterize the positions of the particles of the system as follows @xmath221 .\ ] ] the @xmath222 are then the non - rotating equilibrium dimensionless position coordinates of the particles . equilibrium corresponds to @xmath223 , @xmath8 , see section [ sec.ellipsoids ] . @xmath220 is a dilatation ( or radial pulsation ) degree of freedom , while @xmath224 corresponds to spheroidal deformations and @xmath225 to triaxial deformations . the velocity of the particles are @xmath226 \\ + r \lambda [ ( \dot \xi+\dot \tau ) x^_i { \mbox{\boldmath$e$}}_x + ( \dot \xi-\dot \tau ) y^_i{\mbox{\boldmath$e$}}_y -2(\xi\dot \xi -\tau\dot\tau ) ( \xi^2- \tau^2)^{-2 } z^_i { \mbox{\boldmath$e$}}_z ] .\end{aligned}\ ] ] an elementary computation shows that the kinetic energy , @xmath227 , becomes @xmath228\end{aligned}\ ] ] here @xmath229 , and @xmath230 . the potential energy is given in eq . ( [ eq.energy.xi.tau ] ) . to get an explicit potential energy we must find some expression for the volume dependent ( pressure ) energy @xmath231 . as a model for this part of the energy we take the simple expression @xmath232 where we must take @xmath233 if the corresponding force is to balance gravity and prevent collapse . putting these together we thus find the total potential energy @xmath234 if we use the definition of @xmath11 in eq . ( [ eq.def.k ] ) we find the expression @xmath235 for the potential energy of the system . combining @xmath236 of eq . ( [ eq.kin.energy.2 ] ) with @xmath151 to the lagrangian @xmath237 we can get the dynamics of this three degree of freedom system exactly by finding and solving the euler - lagrange equations of the system . here we will instead assume small oscillations and make a normal mode analysis . we first taylor expand @xmath151 to quadratic terms around @xmath238 and then solve for the zeros of the first derivatives of this quadratic to get the position of the minimum . the result is @xmath239 we now introduce new variables , @xmath240 through @xmath241 then , of course , @xmath242 . in the kinetic energy we replace @xmath243 in the coefficients with the equilibrium positions @xmath244 and expand to first order in @xmath11 . in the potential energy we make the same replacement in the quadratic taylor expansion , expand to first order in @xmath11 and throw away constant terms . the result is that @xmath245\ ] ] and that @xmath246 .\ ] ] only the degrees of freedom @xmath247 and @xmath248 are coupled and only when @xmath249 . for @xmath250 the spheroidal @xmath248-mode and the triaxial @xmath6-mode are degenerate ( have the same frequency ) . the value @xmath251 is special since for that value all three modes have the same frequency . for the @xmath247-mode pressure is restoring force in the contracting phase of the motion and gravity in the expanding phase . for the other two modes , @xmath252 , gravity alone acts to restore the spherical minimum . if we put @xmath253 eq . ( [ eq.def.k ] ) shows that @xmath254 , and eq.([eq.maclaurin.approx.eq ] ) , with @xmath255 , that @xmath256 . for @xmath257 we then get the approximate eigen frequencies @xmath258 to first order in @xmath11 . from this one easily finds the following first order results @xmath259,\\ t_{\tau } = \frac{2\pi}{\omega_{0 } } \left[1+\frac{1}{2}\left(1+\frac{16}{15n}\right)\epsilon\right],\end{aligned}\ ] ] for the corresponding periods . the free radial oscillation mode for the earth is known @xcite to have @xmath260min = @xmath261s . this means that one can calculate @xmath262 and find it to be @xmath263 for the earth . since the equilibrium radius of the rotating earth is @xmath264 eq.([min.pos , lambda ] ) , and a small calculation , now shows that earth s mean radius is @xmath265 m larger due to rotation compared to the non - rotating case . the fairly large @xmath262 also shows that the @xmath266 mode periods are essentially independent of @xmath262 . for the earth one finds that the @xmath6-mode has period @xmath267 min . due to the rotational splitting , which is given by @xmath268 the @xmath248-mode is @xmath269 seconds shorter . the longest observed free oscillation period of the earth has period @xmath270min ( see udas @xcite ) . the main explanation for the discrepancy is probably that the model neglects elastic forces . though not quantitatively reliable from this point of view the model has the advantages of showing in a simple way how rotational splitting of the modes arise and the order of magnitude of such splittings . some apparently new results relating to the classic theory of the figure of rotating bodies have been presented . the basic model , a point mass at the center of a homogeneous fluid , is characterized by their mass ratio , and interpolates between the limits of an ellipsoidal homogeneous fluid and a body dominated by a small central mass concentration . it allows simple analytic treatment but is still flexible enough to correctly describe the essential hydrostatics of real rotating planets as well as stars . such models are always useful , especially when one wishes to analyze , compare , and understand large numbers of observational data . in spite of the simplicity there is no perturbation order to which the results are valid ; the nonlinearity of the basic equations can be retained . to be more precise eq . ( [ eq.maclaurin.approx ] ) shows that the formulas are valid to seventh order in the eccentricity ( within the basic model with its simplified mass distribution ) . as demonstrated by the numerical experiments on jupiter and saturn data , this is essential . in fact table [ table ] indicates that the geometric oblateness of the surface equipotential surface of mars and uranus determined from observed @xmath2 and @xmath1 values using ( [ eq.maclaurin.approx.relation.epsilon ] ) probably are more reliable than current observational data . finally the dynamics of the model reveals the essentials of the coupling and rotational splitting of the most basic free oscillations of a planet without recourse to expansion in spherical harmonics . + + acknowledgement * constructive comments from dr . john d. anderson on a previous version of this manuscript are gratefully acknowledged . the first row for each planet gives literature @xcite data . these are observational except @xmath12 which are based on theoretical calculations . the second row for each planet gives the corresponding calculated values as given by formulas ( [ eq.maclaurin.approx.relation.q ] ) - ( [ eq.eps.e2 ] ) in such a way that for each pair of observational @xmath2 , @xmath1 and @xmath16 the missing third is calculated . the moment of inertia is calculated from observational @xmath2 and @xmath1 using formula ( [ eq.j2.expression ] ) , @xmath271 , with @xmath206 from formula ( [ eq.maclaurin.approx.relation.epsilon ] ) . + 3 . the third row for jupiter and saturn gives data calculated in a similar way to that in the second row except that the first order formula ( [ eq.j2.alpha.epsilon ] ) has been used to find the third value from two observational values . + 4 . for the sun @xmath0 , @xmath1 and @xmath16 have been calculated from an observation based @xmath2-value discussed in the text and a theoretical @xmath12 @xcite , using formulas ( [ eq.ellipt.kappa2 ] ) , ( [ eq.j2.kappa2 ] ) and @xmath272 respectively .	a point mass at the center of an ellipsoidal homogeneous fluid is used as a simple model to study the effect of rotation on the shape and external gravitational field of planets and stars . maclaurin s analytical result for a homogenous body is generalized to this model . the absence of a third order term in the taylor expansion of the maclaurin function leads to further simple but very accurate analytical results connecting the three observables : oblateness ( @xmath0 ) , gravitational quadrupole ( @xmath1 ) , and angular velocity parameter ( @xmath2 ) . these are compared to observational data for the planets . the moments of inertia of the planets are calculated and compared to published values . the oblateness of the sun is estimated . oscillations near equilibrium are studied within the model . + +
quantum physics is a technically difficult and abstract subject . @xcite the subject matter makes instruction quite challenging and students perpetually struggle to master basic concepts . @xcite here i discuss the development and evaluation of quantum interactive learning tutorials ( quilts ) that help advanced undergraduate students learn quantum mechanics . quilts are designed to create an active learning environment where students have an opportunity to confront their misconceptions , interpret the material learned , draw qualitative inferences from quantitative tools learned in quantum mechanics and build links between new material and prior knowledge . they are designed to be easy to implement regardless of the lecturer s teaching style . a unique aspect of quilts is that they are research - based , targeting specific difficulties and misconceptions students have in learning various concepts in quantum physics . @xcite they often employ computer - based visualization tools @xcite to help students build physical intuition about quantum processes and keep students consistently engaged in the learning process by asking them to predict what should happen in a particular situation , and then providing appropriate feedback . they attempt to bridge the gap between the abstract quantitative formalism of quantum mechanics and the qualitative understanding necessary to explain and predict diverse physical phenomena . they can be used in class by the instructors once or twice a week as supplements to lectures or outside of the class as homework or as self - study tool by students . the quilts use a learning cycle approach @xcite in which students engage in the topic via examples that focus their attention , explore the topic through facilitated questioning and observation , explain what they have learned with instructor facilitating further discussion to help refine students understanding and extend what they have learned by applying the same concepts in different contexts . the guidance provided by the tutorials is decreased gradually and students assume more responsibility in order to develop self - reliance . in addition to the main tutorial , quilts often have a warm - up " component and a tutorial homework " . students work on the warm - up " component of a quilt at home before working on the main tutorials in class . these warm - ups typically review the prior knowledge necessary for optimizing the benefits of the main tutorial related to that topic . the tutorial homework " associated with a quilt can be given as part of their homework to reinforce concepts after students have worked on the main tutorial . the tutorial homework helps students apply the topic of a particular quilt to many specific situations to learn about its applicability in diverse cases and learn to generalize the concept appropriately . we design a pre - test and post - test to accompany each quilt . the pre - test assesses students initial knowledge before they have worked on the corresponding quilt , but typically after lecture on relevant concepts . the quilt , together with the preceding pre - test , often make students difficulties related to relevant concepts clear not only to the instructors but also to students themselves . the pre - test can also have motivational benefits and can help students better focus on the concepts covered in the quilt that follows it . pre-/post - test performances are also useful for refining and modifying the quilt . an integral component of the quilts is the adaptation of visualization tools for helping students develop physical intuition about quantum processes . a visualization tool can be made much more pedagogically effective if it is embedded in a learning environment such as quilt . a simulation , preceded by a prediction and followed by questions , can help students reflect upon what they visualized . such reflection can be useful for understanding and remembering concepts . they can also be invaluable in helping students better understand the differences between classical and quantum concepts . we have adapted simulations from a number of sources as appropriate @xcite including the open source java simulations developed by belloni and christian . @xcite some of the quilts , e.g. , the quilt on double - slit experiment which uses simulations developed by klaus muthsam , @xcite are also appropriate for modern physics courses . the double - slit quilt uses simulations to teach students about the wave nature of particles manifested via the double slit experiment with single particles , the importance of the phase of the probability amplitude for the occurrence of interference pattern and the connection between having information about which slit a particle " went through ( which - path " information ) and the loss of interference pattern . for the quilts based on simulations , students must first make predictions about what and why they expect a certain thing to happen in a particular situation before exploring the relevant concepts with the simulations . for example , students can learn about the stationary states of a single particle in various types of potential wells . students can change the model parameters and learn how those parameters affect stationary states and the probability of finding the electron at a particular position . they can also take various linear combinations of stationary states to learn how the probability of finding the electron at a particular position is affected . they can calculate and compare the expectation values of various operators in different states for a given potential . they can also better appreciate why classical physics may be a good approximation to quantum physics under certain conditions . students can also develop intuition about the differences between bound states and scattering states by using visual simulations . guided visualization tools can also help students understand the changes that take place when a system containing one particle is extended to many particles . @xcite similar to the development of tutorials for introductory and modern physics , @xcite the development of each quilt goes through a cyclical iterative process . preliminary tutorials are developed based upon common difficulties in learning a particular topic , @xcite and how that topic fits within the overall structure of quantum mechanics . the preliminary tutorials are implemented in one - on - one interviews with student volunteers , and modifications are made . these modifications are essential for making the quilts effective . after such one - on - one implementation with at least half a dozen students , tutorials are tested and evaluated in classroom settings and refined further . working through quilts in groups is an effective way of learning because formulating and articulating thoughts can provide students with an opportunity to solidify concepts and benefit from one another s strengths . it can also provide an opportunity to monitor their own learning because mutual discussions can help students rectify their knowledge deficiencies . students typically finish a quilt at home if they can not finish it in class and take the post - test associated with it individually in the following class for which no help is provided . below , we briefly discuss case studies related to the development and evaluation of three quilts on time development of wave function , uncertainty principle , and mach - zehnder interferometer . the development of each quilt starts with a careful analysis of the difficulties students have in learning related concepts . after the preliminary development of the tutorials and the pre-/post - tests associated with them , we conduct one - on - one 1.5 hour interviews with 6 - 7 student volunteers for each tutorial using a think - aloud protocol . @xcite in this protocol , students are asked to work on a tutorial while talking aloud so that we could follow their thought processes . hints are provided as appropriate . these individual interviews provide an opportunity to probe students thinking as they work through a tutorial and gauge the extent to which students are able to benefit from them . after each of these interviews , the tutorials are modified based upon the feedback obtained . then , they are administered in the classroom and are modified further based upon the feedback . table 1 shows the performance on the pre-/post - test of advanced undergraduate students in a quantum mechanics course on the last version . the pre - test was given after traditional instruction on relevant concepts but before the tutorial . below we summarize each tutorial and discuss student performance on the case - study . we note that the pre - test and post - test for a quilt were not identical but often had some identical questions . one difficulty with the time development of wave functions stems from the fact that many students believe that the only possible wave functions for a system are stationary states . @xcite since the hamiltonian of a system governs its time development , we may expand a non - stationary state wave function @xmath0 at the initial time @xmath1 in terms of the stationary states and then multiply appropriate time dependent phase factors @xmath2 with each term ( they are in general different for different stationary states because the energies @xmath3 are different ) to find the wave function @xmath4 at time @xmath5 . students often append an overall time - dependent phase factor even if the wave function is in a linear superposition of the stationary states . @xcite to elicit this difficulty , the pretest of this quilt begins by asking students about the time dependence of a non - stationary state wave function for an electron in a one - dimensional infinite square well . if the students choose an overall phase factor similar to that for a stationary state , they are asked for the probability density , i.e. , the absolute square of the wave function . as noted above , when a non - stationary state is expanded in terms of stationary states , the probability density at time @xmath5 , @xmath6 , is generally non - stationary due to a different time - dependent phase factor in each term . if students incorrectly choose that the wave function is time - independent even for a non - stationary state , arguing that overall phase factors cancel out , the tutorial asks them to watch the simulations for the time evolution of the probability densities . simulations for this quilt are adapted from the open source physics simulations developed by belloni and christian . @xcite these simulations are highly effective in challenging students beliefs . students are often taken aback when they find that the probability density oscillates back and forth for the non - stationary state . figure 1 shows snapshots adapted in quilt from an open source physics simulation by belloni and christian for the probability density for a non - stationary state wave function for a one - dimensional harmonic oscillator well . in the actual simulation , students watch the probability density evolve in time . when students observe that the probability density does not depend on time for the stationary - state wave function but depends on time for the non - stationary - state wave function , they are challenged to resolve the discrepancy between their initial prediction and observation . in our model , this is a good time to provide students guidance and feedback to help them build a robust knowledge structure . students then work through the rest of the quilt which provides appropriate support and helps solidify basic concepts related to time development . students respond to time development questions with stationary and non - stationary state wave functions in problems involving different potential energies ( e.g. , harmonic oscillator , free - particle etc . ) to reinforce concepts , and they receive timely feedback to build their knowledge structure . for each case , they check their calculations and predictions for the time - dependence of the probability density in each case with the simulations . within an interactive environment , they learn that the hamiltonian governs the time development of the system , and that the eigenstates of the hamiltonian are special with regards to the time evolution of the system . they learn that not all possible wave functions are stationary - state wave functions , and they learn about the difference between the time - independent and time - dependent schroedinger equation . table 1 shows that in the case study in which nine students took both the pre-/post - tests , the average student performance improved from @xmath7 to @xmath8 after working on the quilt . as discussed earlier , the most common difficulty on the pre - test was treating the time evolution of non - stationary states as though those states were stationary states . moreover , two students who were absent on the day the pre - test and tutorial were administered in the class but were present for the post - test in the following class obtained @xmath9 and @xmath10 on the post - test respectively . the quilt on the uncertainty principle contains three parts with increasing levels of sophistication . depending upon the level of students , the instructors may choose to use only one or all parts . the first part of this quilt helps students understand that this fundamental principle is due to the wave nature of particles . with the help of the de broglie relation , the quilt helps students understand that a sinusoidal extended wave has a well - defined wavelength and momentum but does not have a well - defined position . on the other hand , a wave pulse with a well defined position does not have a well defined wavelength or momentum . students gain further insight into the uncertainty principle in the second part of the quilt by fourier transforming the position - space wave function and noticing how the spread of the position - space wave function affects its spread in momentum space . computer simulations involving fourier transforms are exploited in this part of the quilt and students fourier transform various position - space wave function with different spreads and check the corresponding changes in the momentum - space wave function . the third part of the quilt helps students generalize the uncertainty principle for position and momentum operators to any two incompatible observables whose corresponding operators do not commute . this part of the quilt also helps students bridge this new treatment with students earlier encounter with uncertainty principle for position and momentum in the context of the spread of a wave function in position and momentum space . the quilt also helps students understand why a measurement of one observable immediately followed by the measurement of another incompatible observable does not guarantee a definite value for the second observable . table 1 shows that the average performance of 12 students who took the last version of the quilt improved from @xmath11 to @xmath12 from pre - test to post - test . in a question that was common for both the pre - test and post - test , students were asked to make a qualitative sketch of the absolute value of the fourier transform of a delta function . they were asked to explain their reasoning and label the axes appropriately . only one student in the pre - test drew a correct diagram . in the post - test , 10 out of 12 students were able to draw correct diagrams with labeled axes and explain why the fourier transform should be a constant extended over all space . also , in the post - test , 10 out of 12 students were able to draw the fourier transform of a gaussian position space wave function and were able to discuss the relative changes in the spread of the position and the corresponding momentum space wave functions . these were concepts they had explored using computer simulations while working on the quilt . similar results were found in individual interviews conducted earlier with other students during the development of the quilt . one of the questions on both the pre-/post - test of this tutorial was the following : + consider the following statements : uncertainty principle makes sense . when the particle is moving fast , the position measurement has uncertainty because you can not determine the particle s position precisely .. it is a blur .... that s exactly what we learn in quantum mechanics .. if the particle has a large speed , the position measurement can not be very precise . " explain why you agree or disagree with the statement . + out of the 12 students who took both pre-/post - tests , 7 students provided incorrect responses on the pre - test . the following are examples of incorrect student responses on the pre - test : 1 . _ i agree ... when p is high it is easy to determine , while x is difficult to determine . the opposite is also true , when p is small it is difficult to determine , while x is easy to determine . _ i agree because when a particle has a high velocity it is difficult to measure the position accurately _ 3 . _ i agree because i know the uncertainty principle to be true . _ when a particle is moving fast , we can not determine its position exactly - it resembles a wave - at fast speed , its momentum can be better determined . _ in comparison , one student provided incorrect response and one did not provide a clear reasoning on the post - test . the goals of this quilt are to review the interference at a detector due to the superposition of light from the two paths of the interferometer . the tutorial adapts a simulation developed by albert huber ( http://www.physik.uni-muenchen.de/didaktik/computer/interfer/interfere.html ) to help students learn the following important quantum mechanical concepts : * interference of a single photon with itself after it passes through the two paths of the mz . * effect of placing detectors and polarizers in the path of the photon in the mz . * how the information about the path along which a photon went ( which - path " information ) destroys the interference pattern . a screen shot from the simulation is shown in figure 2 . students were given the following information about the setup : the basic schematic setup for the mach - zehnder interferometer ( mz ) used in this quilt is as follows ( see figure 3 ) with changes made later in the tutorial , e.g. , changes in the position of the beam splitters , incorporation of polarizers , detectors or a glass piece , to illustrate various concepts . all angles of incidence are @xmath13 with respect to the normal to the surface . for simplicity , we will assume that light can only reflect from one of the two surfaces of the identical half - silvered mirrors ( beam splitters ) @xmath14 and @xmath15 because of anti - reflection coatings . the detectors @xmath16 and @xmath17 are detectors located symmetrically with respect to the other components of the mz as shown . the photons originate from a monochromatic coherent point source . assume that the light through both the @xmath18 and @xmath19 paths travels the same distance in to reach each detector . + in this quilt , students first learn about the basics of phase changes that take place as light reflects or passes through different beam splitters and mirrors in the mz setup by drawing analogy with reflected or transmitted wave on a string with fixed or free boundary condition at one end . then , students use the simulation to learn that a single photon can interfere with itself and produce interference pattern after it passes through both paths of the mz . students explore and learn using simulations that which - path " information is obtained by removing @xmath15 or by placing detectors or polarizers in certain locations . later in the tutorial , point detector @xmath16 is replaced with a screen . table 1 shows that the average performance of 12 students who took the last version of the mz quilt improved from @xmath20 to @xmath12 from pre - test to post - test . moreover , all but one of the 12 students in the post - test obtained perfect scores on the following three questions ( correct options ( c ) , ( b ) , and ( b ) respectively ) that were similar ( but not necessarily identical to ) the kinds of setups they had explored using the simulation within the guided quilt approach : 1 . if you insert polarizers 1 and 2 ( one with a horizontal and the other with a @xmath13 transmission axis ) as in the figure 4 , how does the interference pattern compare with the case when the two polarizers have orthogonal transmission axis ? + ( a ) the interference pattern is identical to the case when polarizers 1 and 2 have orthogonal axes . + ( b ) the interference pattern vanishes when the transmission axes of polarizers 1 and 2 are horizontal and @xmath13 . + ( c ) an interference pattern is observed , in contrast to the case when polarizers 1 and 2 were orthogonal to each other . + ( d ) no photons reach the screen when the transmission axes of polarizers 1 and 2 are horizontal and @xmath13 . if you insert polarizer 1 with a horizontal transmission axis and polarizer 2 ( between the second beam splitter and the screen ) with a @xmath13 transmission axis ( figure 5 ) , how does the interference pattern compare with the case when only polarizer 1 was present ? + ( a ) the interference pattern is identical to the case when only polarizer 1 was present . + ( b ) the intensity of the interference pattern changes but the interference pattern is maintained in the presence of polarizer 2 . + ( c ) the interference pattern vanishes when polarizer 2 is inserted but some photons reach the screen . + ( d ) an interference pattern reappears that was absent when only polarizer 1 was present . if you insert polarizer 2 with a @xmath13 transmission axis between the second beam splitter and the screen ( figure 6 ) , how does the interference pattern compare with the case when polarizer 2 was not present ? + ( a ) the interference pattern is unchanged regardless of the presence of polarizer 2 because all interference effects occur before beam splitter 2 . + ( b ) the intensity of the interference pattern decreases but the interference pattern is maintained even in the presence of polarizer 2 . + ( c ) the intensity of the interference pattern increases in the presence of polarizer 2 . + ( d ) the interference pattern vanishes when polarizer 2 is inserted but some photons reach the screen . a survey of 12 students whose pre-/post - test data is presented in table 1 was given to assess the effectiveness of quilts from students perspective . below we provide the questions and student responses : + * please rate the tutorials for their overall effectiveness where 1 means totally ineffective and 5 means very effective . + in response to this question , no student chose 1 or 2 , one student chose 3 , one chose 3.5 , three chose 4 , one 4.5 and six chose 5 . * how often did you complete the tutorial at home that you could not complete during the class ? ( 1 ) never , ( 2 ) less than half the time , ( 3 ) often , ( 4 ) most of the time , ( 5 ) always . + in response to this question , no student chose ( 1 ) , one student chose ( 2 ) , two students chose ( 3 ) , 6 chose ( 4 ) , and 3 chose ( 5 ) . * how often were the hints / solutions provided for the tutorials useful ? ( 1 ) never , ( 2 ) less than half the time , ( 3 ) often , ( 4 ) most of the time , ( 5 ) always . + in response to this question , no student chose ( 1 ) or ( 2 ) , 2 students chose ( 3 ) , 5 chose ( 4 ) and 5 chose ( 5 ) . * is it more helpful to do the tutorials in class or would you prefer to do them as homework ? please explain the advantages and disadvantages as you see it . + in response to this question , 10 students felt that doing them in class was more useful . the students who preferred doing them in class often noted that the tutorials focused on improving their conceptual understanding which was best done via group discussion and hence in class . they appreciated the fact that any questions they had could be discussed and they benefited from the reasoning provided by their peers and instructor . the few students who preferred doing them at home felt that more time and effort will go into them if they did them at home . * how frequently should the tutorials be administered in the class ( e.g. , every other class , once a week , once every other week ) ? explain your reasoning . + a majority of students liked having the tutorials once a week . this frequency was considered to be the best by some students who felt that the concepts learned in the tutorials made it easier for them to understand the textbook and homework problems later in the week and integrate the material learned . others felt that once a week was the best because tutorials helped them focus on concepts that were missed in lectures , book , and student / teacher conversation . * do you prefer a multiple - choice or open - ended question format for the tutorial questions ? explain your reasoning . + students in general seemed to like the questions that were in the multiple - choice format but most of them also appreciated the open - ended questions . some students noted that the multiple - choice questions helped focus their attention on important issues and common difficulties and misconceptions while the open - ended questions stimulated creative thought . some students felt that multiple - choice format may be better for the warm - up " tutorial done at home and the open - ended questions may be better for the main tutorial done in the class . some students felt that a mix of the two types of questions was best because the multiple - choice format was a good way to get the fundamental concepts across and the open - ended questions gave them an opportunity to apply these concepts and deepen their understanding of the concepts . we have given an overview of the development of quilts and discuss the preliminary evaluation of three quilts using pre-/post - tests in the natural classroom setting . quilts adapt visualization tools to help students build physical intuition about quantum processes . they are designed to help undergraduates sort through challenging quantum mechanics concepts . they target misconceptions and common difficulties explicitly , focus on helping students integrate qualitative and quantitative understanding , and learn to discriminate between concepts that are often confused . they strive to help students develop the ability to apply quantum principles in different situations , explore differences between classical and quantum ideas , and organize knowledge hierarchically . their development is an iterative process . during the development of the existing quilts , we have conducted more than 100 hours of interviews with individual students to assess the aspects of the quilts that work well and those that require refinement . quilts naturally lend themselves to dissemination via the web . they provide appropriate feedback to students based upon their need and are suited as an on - line learning tool for both undergraduates ( and beginning graduate students ) in addition to being suitable as supplements to lectures for a one or two - semester undergraduate quantum mechanics courses . we are very grateful to mario belloni and wolfgang christian for their help in developing and adapting their open source physics simulations for quilts . we also thank albert huber for mach zehnder interferometer simulation and to klaus muthsam for the double slit simulation . we thank all the faculty who have administered different versions of quilts in their classrooms . + e. j. galvez , c. h. holbrow , m. j. pysher , j. w. martin , n. courtemanche , l. heilig and j. spencer , _ interference with correlated photons : five quantum mechanics experiments for undergraduates _ , am . j. phys . * 73 * , 127 - 140 ( 2005 ) . c. singh , _ improving student understanding of quantum mechanics _ , proceedings of the phys . conference , salt lake city , ut , edited by p. heron , l. mccullough , j. marx , 69 - 72 , ( aip , melville ny , 2005 ) . c. singh , helping students learn quantum mechanics for quantum computing , " proc , phys . conference , syracuse , ny , edited by l. mccullough , l. hsu and p. heron , 42 - 45 ( aip , melville ny , 2006 ) . m. belloni and w. christian , _ physlets for quantum mechanics _ , comp . eng . 5 , 90 ( 2003 ) ; m. belloni , w. christian and a. cox , _ physlet quantum physics _ , pearson prentice hall , upper saddle river , nj , ( 2006 ) . mach - zehnder simulation was adapted from http://www.physik.uni-muenchen.de/didaktik/computer/interfer/interfere.html and double - slit simulation was developed by klaus muthsam ( [email protected] ) . the original spins program was written by daniel schroeder and thomas moore for the macintosh and was ported to java by david mcintyre of oregon state university and used as part of the paradigms project . both of these versions were , and remain , open source .	we discuss the development and evaluation of quantum interactive learning tutorials ( quilts ) , suitable for one or two - semester undergraduate quantum mechanics courses . quilts are based upon investigation of student difficulties in learning quantum physics . they exploit computer - based visualization tools and help students build links between the formal and conceptual aspects of quantum physics without compromising the technical content . they can be used both as supplements to lectures in the classroom or as a self - study tool . address = department of physics , university of pittsburgh , pittsburgh , pa , 15213
the term `` master equation '' ( me ) used in this work means an equation of motion for the reduced density operator @xmath0 of subsystem @xmath1 which interacts with ( usually much larger ) another subsystem @xmath2 . experiences gathered during quite a few years of lecturing indicate that students find it difficult to understand the derivation and consequences of me . the literature sources ( at least those known to us ) are usually quite brief and not easy to follow . the aim of this paper is to give a full presentation of the so - called microscopic derivation of me together with detailed discussion of the underlying assumptions and approximations . hence , this work does not bring any original material . it is just a tutorial which , hopefully , will help the students to obtain a better grasp of the ideas and concepts leading to one of important theoretical methods used in a variety of physical problems . the literature of the subject is quite rich . entering the term `` master equation '' into _ google scholar _ returns about 30 000 links . it seems virtually impossible to give an extensive bibliography . therefore , we concentrate only on several works which were essential in the preparation of this paper . in order to present the me technique , we first give the basic ideas which help , so to speak , to set the scene for further developments . we assume that the reader is familiar with fundamental concepts of quantum mechanics ( given , for example , in the main chapters of the first volume of the excellent book by cohen - tannoudji _ et al _ @xcite ) . to start with , we recall that the state of the quantum system is , in the majority of practical applications , given by the density operator . this concept is introduced and discussed in virtually all handbooks on quantum mechanics , hence we mention only the basic facts which will be needed here . namely , density operator for any physical system must have three essential properties [ ii01 ] @xmath3 the last inequality means that @xmath4 for any state @xmath5 from the hilbert space of the states of the given physical system . these properties may be phrased in terms of the eigenvalues @xmath6 of the density operator . namely , @xmath7 , @xmath8 , and @xmath9 . for a closed system with hamiltonian @xmath10 , the evolution of the corresponding density operator is given by the von neumann equation @xmath11 , \label{ii03}\ ] ] which has a well - known solution @xmath12 where @xmath13 is a suitably chosen initial condition . such an evolution is unitary @xcite and obviously preserves all the necessary properties of the density operator . in such a case everything is conceptually clear although necessary calculation may be quite involved or even requiring some approximate computational methods . the problem arises when we deal with a bipartite system @xmath14 , consisting of two interacting subsystems @xmath1 and @xmath2 . let us briefly outline the physical situation . we assume that the whole system @xmath14 is closed and the total hamiltonian is written as @xmath15 where @xmath16 and @xmath17 are free , independent hamiltonians of each of the subsystems @xmath18 and @xmath19 . @xmath20 is the hamiltonian describing the interaction between two parts . let @xmath21 denote the density operator of the total system . then , @xmath21 evolves according to the von neumann equation ( identical as ) @xmath22 , \label{ii06}\ ] ] now , one may ask , what is the problem ? the point is that we are , in fact , interested only in the subsystem @xmath1 . subsystem @xmath2 , for this reason or other , is considered irrelevant , although the @xmath1-@xmath2 interaction certainly affects the evolution of @xmath1 . moreover , in many practical cases , subsystem @xmath2 is much larger ( with many more degrees of freedom ) and virtually inaccessible to direct measurements . frequently , @xmath2 is a reservoir @xcite and plays the role of environment upon which we have neither control nor influence . this may be very important in the context of quantum information theory @xcite when the relevant subsystem is disturbed by the surroundings . moreover , problems of decoherence and irreversibility are intrinsically connected with the effects occurring in a subsystem influenced by an external reservoir ( see @xcite ) . we shall not discuss these problems but focus on master equation technique . in the view of these brief remarks the question is , how to extract useful information on @xmath1 from the general von neumann equation . we stress that we are interested only in system @xmath1 , so we need to find the reduced density operator @xmath23 the aim is , therefore , twofold : * extract the evolution of @xmath0 from eq . for the entire system @xmath24 . * do it in a way which guarantees that properties of @xmath0 ( as of any density operator ) are preserved for any moment of time . the solution to the stated problems is found in the so - called master equation technique . there are two , conceptually different but complementary , approaches to me . the first one uses mathematically rigorous methods . such an approach is presented , for example , in refs . @xcite ( see also the references given in these books ) . rigorous mathematics is obviously very important from the fundamental point of view . mathematical theorems prove that me for the reduced density operator of subsystem @xmath1 follows from general von neumann equation and indeed preserves properties . in refs . @xcite it is shown that it is so , when me attains the so - called standard ( lindblad - gorini - kossakowski - sudarshan ) form ( ref . @xcite , p.8 , eq . ( 38 ) ; @xcite , eq . ( 3.63 ) and @xcite , eq . ( 78 ) ) @xmath25 ~+~ \sum_{ij } a_{ij } \biggl ( \ , f_{i } \ , \rho_{a}(t ) \ , f_{j}^{\dagger } ~-~ { { \textstyle \frac{1}{2}}}\bigl [ f_{j}^{\dagger } f_{i } , \ ; \rho_{a}(t ) \bigr]_{+ } \ , \biggr ) . \label{ii10}\ ] ] the first term leads the unitary ( hamiltonian ) evolution , while the second one is sometimes called a dissipator @xcite . operators @xmath26 constitute a basis in the space of operators for @xmath1 subsystem ( see @xcite ) . finally , @xmath27 is a positive definite hermitian matrix . here , we do not go into the details of the derivation of the standard form of the master equation , we refer the reader to the mentioned references ( see also @xcite ) . quite interesting derivation of standard me is given by preskill @xcite . his presentation is somewhat heuristic but certainly worth reading , especially if one is interested in connection with quantum information theory , quantum channels etc . it is not , however , our aim to pursue formal mathematical issues . our intentions are quite practical , so the reader may ask , why do we speak about the standard form of me . the reason is as follows . the second approach to the stated problems is via `` microscopic derivations '' . this occurs when we need to consider a specific physical situation when interacting systems @xmath1 and @xmath2 are known and well - defined . then , we want to construct the corresponding me equation of motion for the reduced density operator @xmath0 . this is less formal and may be mathematically uncertain . if the microscopic derivation yields an equation in the standard form , we can say that the aim is achieved , because standard form ensures the preservation of the necessary properties of reduced density operator @xmath0 . hence , both approaches are complementary . formal but rigorous mathematical methods lead to standard form of me which must be matched by equations obtained through microscopic derivation . all of the already mentioned references give ( usually brief ) account on the microscopic derivation of me . these presentations seem to be difficult for students who are not acquainted with the subject and who seek the necessary introduction . perhaps the most extensive microscopic derivation is given by cohen - tannoudji , dupont - roc and grynberg @xcite . this latter presentation is somewhat heuristic and , as it seems to us , leaves some nuances unexplained . there is , however , one more drawback . namely , cohen - tannoudji _ et al _ do not compare their me with the standard form . therefore , essential question of positivity preservation remains untouched . the derivation given here uses the concepts which can be found in the all references . nevertheless , we are most indebted to cohen - tannoudji _ et al_. their approach and especially their discussion of the time scales and employed approximations strongly influenced our tutorial . we apologize if , at some places , we do not give proper references . too many ones can distract the reader , which does not lessen our debt to all authors of cited literature . the scheme of this paper is summarized in the _ contents _ , hence we feel no need to repeat it . we try to be as clear and as precise as possible . we focus our attention on the microscopic derivation of me , leading to the standard form . some issues are postponed to _ auxiliary _ section , so that the main flow of derivation is not broken by additional comments , which can safely be given later . we hope , that the students who need a primer in the subject , will find our work useful and informative . the readers are invited to comment , so that the next version ( if such a need arises ) will be improved and , more readable . we consider a physical system which consists of two parts @xmath1 and @xmath2 . we are interested only in what happens in part @xmath1 which is usually much smaller than part @xmath2 . the latter one we will call a reservoir ( environment ) . we assume that the entire system @xmath24 is closed . then , its hamiltonian is specified as in eq . some additional assumptions concerning both subsystems will be introduced when necessary . a previously , let @xmath21 denote the density operator of the whole system @xmath24 . the evolution of this operator is governed by von neumann equation . our main aim is to find the corresponding equation of motion for the reduced density operator @xmath28 for the subsystem @xmath1 . our starting point is provided by von neumann equation which , after the transformation to the interaction picture , reads @xmath29 , \label{me05}\ ] ] where , we obviously denoted @xmath30 with @xmath31 given in eq . . reduction of the density operator ( as in ) is preserved in the interaction picture ( see _ auxiliary _ sections ) @xmath32 formal integration of eq . yields the following expression @xmath33 , \label{me07}\ ] ] which gives the density operator at a later moment @xmath34 , while the initial one at a moment @xmath35 is assumed to be known . iterating further and denoting @xmath36 we obtain , ( similarly as in ref.@xcite ) @xmath37 \nonumber \\ & \hspace{8 mm } ~+~ \left ( \frac{1}{i\hbar } \right)^{2 } \int\limits_{t}^{t+\delta t } dt_{1 } \int_{t}^{t_{1 } } dt_{2 } ~\bigl [ \ : \widetilde{v}_{ab}(t_{1 } ) , \ ; \bigl [ \ : \widetilde{v}_{ab}(t_{2 } ) , ~{\widetilde{\varrho}}_{ab}(t ) \ : \bigr ] \nonumber \\[2 mm ] & \hspace{-8 mm } ~+~ \left ( \frac{1}{i\hbar } \right)^{3 } \int\limits_{t}^{t+\delta t } dt_{1 } \int_{t}^{t_{1 } } dt_{2 } \int_{t}^{t_{2 } } dt_{3 } ~\bigl [ \ : \widetilde{v}_{ab}(t_{1 } ) , \ ; \bigl [ \ : \widetilde{v}_{ab}(t_{2 } ) , \bigl [ \ : \widetilde{v}_{ab}(t_{3})~{\widetilde{\varrho}}_{ab}(t_{3 } ) \ : \bigr ] . \label{me12 } \end{aligned}\ ] ] higher order iterations will contain fourfold , etc . , integrals and commutators . let us note that in the last term we have time ordering @xmath38 . the above equation is rigorous , no approximations have been made . weak - coupling approximation ( discussed in all refs.@xcite@xcite ) , consists in retaining the terms up to the second order in interaction hamiltonian . higher order terms are then neglected . thus , we remain with @xmath39 \nonumber \\ & \hspace{14 mm } ~+ \left ( \frac{1}{i\hbar } \right)^{2 } \int\limits_{t}^{t+\delta t } dt_{1 } \int_{t}^{t_{1 } } dt_{2 } ~\bigl [ \ : \widetilde{v}_{ab}(t_{1 } ) , \ ; \bigl [ \ : \widetilde{v}_{ab}(t_{2 } ) , ~{\widetilde{\varrho}}_{ab}(t ) \ : \bigr ] . \label{me16 } \end{aligned}\ ] ] alternatively , we can say that the obtained equation is valid in the second - order perturbation theory . such an approximation requires a justification . this will be presented in the _ auxiliary _ sections , now we focus on further steps of the derivation . reduction of the operator @xmath40 in the left hand side of poses no difficulties . tracing over the reservoir variables ( subsystem @xmath19 ) we obtain @xmath41 \nonumber \\ & \hspace{14 mm } ~+~ \left ( \frac{1}{i\hbar } \right)^{2 } \int\limits_{t}^{t+\delta t } dt_{1 } \int_{t}^{t_{1 } } dt_{2 } ~{\mathrm{tr}}_{b } \bigl [ \ : \widetilde{v}_{ab}(t_{1 } ) , \ ; \bigl [ \ : \widetilde{v}_{ab}(t_{2 } ) , ~{\widetilde{\varrho}}_{ab}(t ) \ : \bigr ] . \label{me16x } \end{aligned}\ ] ] this expression has certain drawback . the the commutators in the right hand side contain full density operator @xmath40 , and not the interesting ( relevant ) reduced one @xmath42 . to proceed , we need some more assumptions and approximations . one more remark seems to be in place . subsequent iterations leading to eq . are rigorous . in equation which is approximate there occurs the operator @xmath21 , taken at the initial moment . the last term in the exact equation contains @xmath43 for moments earlier than the current moment @xmath34 , but later than the initial instant @xmath35 . this means that we neglect the influence of the `` history '' on the present moment . we shall return to the discussion of this point . the key role in our consideration is played by the assumption that there are two distinct time scales @xcite . the first one is specified by @xmath44 typical time during which the internal correlations in the reservoir @xmath2 exist . this will discussed in more detail later . here we will only say that @xmath44 is such a time , that when it elapses , the state of the reservoir is practically independent of its initial state . the second scale is provide by time @xmath45 . it characterizes the evolution ( changes ) of the operator @xmath42 which is ( are ) due to the interaction with reservoir , and which may be specified by the relation @xmath46 time @xmath45 may be called the characteristic relaxation time of subsystem @xmath18 . let us note that we are speaking about interaction the interaction picture we employ is thus , particularly useful . we make no statements about the rate of the free evolution of @xmath47 ( in the schrdinger picture ) , which is governed by hamiltonian @xmath16 . usually , the characteristic times of free evolution ( the times of the order of @xmath48 ) are typically much shorter that @xmath45 describing the influence of the interaction between subsystems @xcite . now , we assume that the introduced time scales satisfy the requirement @xmath49 we have a fast scale ( small @xmath44 ) determining the decay of correlations within the reservoir and the second much slower scale defined by relatively long relaxation time @xmath45 , characterizing the interaction between the two parts of the entire physical system . this may be phrased differently . we have assumed that the interaction is weak . let @xmath50 denote the average `` strength '' of this interaction . uncertainty principle states that @xmath51 the condition @xmath52 implies that @xmath53 in other words , we can say that spectral widths are the reciprocals of characteristic times , so the condition @xmath52 means that the spectral width of the reservoir energies must be much larger than the spectral width of the interaction between subsystem @xmath18 with reservoir . further discussion and justification of our approximations is postponed to _ auxiliary _ sections . here we focus on the derivation of the master equation . the adopted assumption @xmath52 allows us to make the approximation which is sometimes called the born one @xcite-@xcite . initial density operator for the whole system @xmath54 can always be written as @xmath55 where @xmath42 and @xmath56 are the reduced density operators for two subsystems . the state of the whole system consists of a factorizable part @xmath57 and the entangled part @xmath58 , describing the interaction - induced correlations between the subsystems . equation gives us the change @xmath59 , hence informs us about changes occurring in the time interval @xmath60 . assumption that @xmath61 allows us to neglect the mentioned correlations . as previously , we postpone the discussion for _ auxiliary _ sections . at present , we assume that @xmath62 by assumption , the reservoir ( environment ) is large , its correlation time is very short , so the reservoir s relaxation is fast . we may say that before any significant changes occur in subsystem @xmath1 , the reservoir would have enough time to reach thermodynamic equilibrium . as it is known from statistical physics such state is given as @xmath63 the quantity @xmath64 is a partition sum @xmath65 states @xmath66 and energies @xmath67 are the eigenstates and eigenvalues of the reservoir hamiltonian , which can be written as @xmath68 . as a consequence we conclude that @xmath69 ~=~ 0 , \label{me23d}\ ] ] so we can say that operator @xmath70 is stationary does not change in time . it is worth noting that we could have reversed the argument . first require stationarity , as expressed by which would entail relations and . moreover , commutation relation implies that the reduced density operator @xmath70 is of the same form both in schrdinger and interaction pictures . due to these remarks operator @xmath56 appearing in eq . is simply replaced by @xmath70 . therefore , in eq . we make the replacement @xmath71 . so , we now have @xmath72 \nonumber \\ & \hspace{1 mm } + \left ( \frac{1}{i\hbar } \right)^{2 } \int\limits_{t}^{t+\delta t } dt_{1 } \int_{t}^{t_{1 } } dt_{2 } ~{\mathrm{tr}}_{b } \bigl [ \ : \widetilde{v}_{ab}(t_{1 } ) , \ ; \bigl [ \ : \widetilde{v}_{ab}(t_{2 } ) , \ ; { \widetilde{\varrho}}_{a}(t ) \otimes { \bar{\sigma}_{b}}\ : \bigr ] . \label{me27 } \end{aligned}\ ] ] the employed simplification facilitates computation of the remaining traces . however , to proceed effectively , we need to specify the interaction hamiltonian . our next assumption concerns the shape of the interaction hamiltonian . it is taken as ( similarly as in the given references ) @xmath74 where @xmath75 are operators which act in the space of the states of subsystem @xmath18 , while operators @xmath76 correspond to space of the reservoir s states . operators appearing in the definition need not be hermitian ( each one separately ) . only the hamiltonian @xmath20 must be hermitian . that is why we have written the second equality . we can say that to each nonhermitian term @xmath77 corresponds the term @xmath78 , and the latter appears in the sum @xmath20 , but with another number . in _ auxiliary _ sections we will argue that it is not any limitation . it is only important that the whole hamiltonian @xmath20 must be hermitian . operators @xmath75 and @xmath76 act in different spaces so they are independent and commute . in the interaction picture we immediately have @xmath79 with @xmath80 rules of hermitian conjugation imply that the conjugate operators transform to interaction picture in the exactly the same manner as the initial ones . we now make one more assumption about reservoir . we have already taken @xmath81 . here , we assume that in the schrdinger picture @xmath82 this assumption easily transforms to interaction picture @xmath83 which follows from commutation relation , cyclic property of trace and . this is rather a simplification , not a restrictive assumption . this will be clarified and explained in _ auxiliary _ sections . ( leading to ) allows us to see that the first term in the me is , in fact , zero . indeed @xmath84 = ~ { \mathrm{tr}}_{b } \biggl [ \ : \sum_{\alpha } \widetilde{a}_{\alpha}(t_{1 } ) \otimes \widetilde{x}_{\alpha}(t_{1 } ) , ~{\widetilde{\varrho}}_{a}(t ) \otimes { \bar{\sigma}_{b}}\ : \biggr ] \nonumber \\ & \hspace{5 mm } = ~\sum_{\alpha } \bigl\ { \ : \widetilde{a}_{\alpha}(t_{1 } ) { \widetilde{\varrho}}_{a}(t ) \ : { \mathrm{tr}}_{b } \bigl [ \widetilde{x}_{\alpha}(t_{1 } ) \ : { \bar{\sigma}_{b}}\bigl ] \nonumber \\ & \hspace{35 mm } ~-~ { \widetilde{\varrho}}_{a } \widetilde{a}_{\alpha}(t_{1 } ) \ : { \mathrm{tr}}_{b } \bigl [ { \bar{\sigma}_{b}}\ : \widetilde{x}_{\alpha}(t_{1 } ) \bigl ] \bigl\ } ~=~ 0 . \label{me40 } \end{aligned}\ ] ] both traces are equal ( cyclic property ) , nevertheless this expression need not be zero , because operators of the @xmath18 system need not commute . if requirement is not fulfilled then the above average may not vanish . assumption and its consequence fortunately give zero , thus the first term of eq . vanishes and we remain with the master equation @xmath85 \right ] . \label{me41a}\ ] ] expanding the commutators is simple . moreover , one easily notices that there are two pairs of hermitian conjugates . hence we have @xmath86 we can now use hamiltonian and perform further transformations . it can be , however , shown @xcite@xcite that this equation does not guarantee that the positivity of the density operator @xmath42 is preserved . it appears that the so - called secular approximation is necessary . to perform it effectively , it is convenient to present the interaction hamiltonian in a somewhat different form . let us write the hamiltonian of the subsystem @xmath1 as @xmath88 states @xmath89 constitute the complete and orthonormal basis in the space of states of the subsystem @xmath1 . the eigenfrequencies @xmath90 may or may not be degenerate . we allow @xmath91 for @xmath92 . at present it suffices that we distinguish different kets @xmath89 solely by their `` _ quantum number _ '' @xmath93 . similarly as in refs.@xcite , we now define the operators @xmath87 via the following relation @xmath94 this representation may be called the decomposition of operator @xmath75 into eigenprojectors of hamiltonian @xmath16 . delta @xmath95 is of the kronecker type , that is @xmath96 in our considerations we allow for nonhermitian operators @xmath75 . hence , definition is augmented by the following one @xmath97 because it is always allowed to interchange the summation indices @xmath98 . we stress that @xmath87 contains bohr frequency @xmath99 , while in @xmath100 we have @xmath101 . the following relation seems to be quite obvious @xmath102 since out of all @xmath103 s one will exactly match @xmath104 . as a consequence we obtain @xmath105 indeed , from definition , relation and due to completeness of states @xmath89 we get @xmath106 relation implies that the interaction hamiltonian ( in schrdinger picture ) can be written as @xmath107 similarly as above we can also show that @xmath108 and @xmath109 using definition we find the operator @xmath110 in the interaction picture @xmath111 because @xmath112 . linking expressions and we write the interaction hamiltonian in the interaction picture @xmath113 before starting to analyze me , let us notice that operators @xmath87 possess some interesting properties @xcite . discussion of these properties is moved to _ auxiliary _ sections . we return to master equation . interaction hamiltonian @xmath114 is taken as in the first part of eq . , while @xmath115 is represented by its hermitian conjugate according to second part of . this gives @xmath116 e^{i\omega\ , ' t_{1 } } a^{\dagger}_{\alpha}(\omega\ , ' ) \otimes \widetilde{x}^{\dagger}_{\alpha}(t_{1 } ) \nonumber \\[2 mm ] & \hspace{-4 mm } -~e^{i\omega\ , ' t_{1 } } a^{\dagger}_{\alpha}(\omega\ , ' ) \otimes \widetilde{x}^{\dagger}_{\alpha}(t_{1 } ) \bigl [ e^{-i\omega t_{2 } } a_{\beta}(\omega ) \otimes \widetilde{x}_{\beta}(t_{2 } ) \bigr ] { \widetilde{\varrho}}_{a}(t ) \otimes { \bar{\sigma}_{b}}\bigr\ } ~+~ \mathbb{h.c}. \label{me60 } \end{aligned}\ ] ] computing tensor products we remember that partial trace is taken only with respect to reservoir variables . moreover , we note that these traces are the same ( cyclic property ) . therefore we denote @xmath117 finally we rewrite the arguments of the exponentials as @xmath118 . . becomes @xmath119 ~+~ \mathbb{h.c}. \label{me64 } \end{aligned}\ ] ] the quantity @xmath120 is called the correlation function of the reservoir . we will briefly discuss its properties . let us focus on the functions defined by the right hand side of eq . , that is @xmath122 these are the functions of two variables and it is not , _ a priori _ , clear that they depend only of the difference @xmath123 . before discussing this fact , let us note that @xmath124 . to prove this relation , we recall that @xmath125 , so that the definition gives @xmath126 now , we will show that the function @xmath127 is indeed a function of the difference of its arguments . the key role plays the fact that the state of the reservoir ( density operator @xmath70 ) is stationary ( does not change in time ) . explicitly using the interaction picture we get @xmath128 the trace is cyclic and @xmath70 commutes with hamiltonianem @xmath17 so we conclude that @xmath129 for two moments of time @xmath130 . reservoir s correlation function effectively depends only on one variable . this fact is denoted by a bar over the symbol of correlation function . thus we write @xmath131 such correlation functions are called stationary . sometimes the concept of stationarity means invariance with respect to time translation . indeed , for arbitrary time @xmath132 we have @xmath133 this property of the correlation functions is a straightforward consequence of the stationarity of reservoir s density operator @xmath70 . preparing this section of our tutorial we greatly benefited from the analogous discussion by cohen - tannoudji _ et al _ @xcite . to large extent we follow their reasoning , trying to elucidate some less obvious points . admitting this , we will refrain from giving multiple references to their work . in master equation one integrates over the triangle @xmath134 which is shown in fig . [ xmerys01 ] . firstly , one computes the integral over @xmath135 in the range from @xmath35 to @xmath136 . this is indicated by thin vertical lines ( at left ) . next , one sums such contributions by integrating over @xmath137 from @xmath35 to @xmath34 . [ 0.6 ] [ 0.6 ] the integrand in contains correlation functions of the reservoir which depend on the difference @xmath123 . we stress that we always have @xmath138 , so that @xmath139 . the integration over the triangle can be performed in another manner . let us consider the geometry ( right graph in fig . [ xmerys01 ] ) . along the diagonal ac we have @xmath140 , so @xmath141 . the straight line @xmath142 has ( in @xmath136 and @xmath143 variables ) the equation @xmath144 , where @xmath145 is fixed , since @xmath146 is the coordinate @xmath143 of the point where the discussed line intersects the axis @xmath143 . then , for the line @xmath147 ( passing through point b ) @xmath145 is also fixed ( by the same argument , as in the case of line @xmath142 ) . on @xmath147 , at the point b we have @xmath148 and @xmath149 . thus , we have @xmath150 . parametr @xmath145 specifies the skew straight lines ( parallel to the diagonal ac ) and passing through triangle abc . [ 0.6 ] integration over the triangle abc is now done as follows . we fix @xmath151 and we move along the segment ac ( see fig.[xmerys02 ] ) . variable @xmath136 runs in the interval from @xmath152 to @xmath34 . so , we first integrate over @xmath137 from @xmath152 to @xmath34 ( along the segment ac ) . next , we integrate over @xmath153 from zero to @xmath60 . in this manner we sum the contributions from all skew segments covering the triangle abc . therefore , we can write @xmath154 while we remember that @xmath155 ( or @xmath156 ) . performing the discussed changes of integration variables in eq . , we get @xmath157 ~+~ \mathbb{h.c}. , \label{me71 } \end{aligned}\ ] ] we recall that the considered time intervals satisfy the requirement @xmath158 ( which will be discussed in detail later ) . if it is true , then the main contribution to the integral over @xmath153 in eq . will come from the region in the neighborhood of @xmath159 . geometrically , this corresponds to a narrow belt which is parallel to the diagonal ac and lies just below it . it follows , that outside this region the reservoir s correlation functions practically vanish ( decay to zero ) . therefore , we will not make any serious error moving the upper limit of integration over @xmath153 to infinity . moreover , since only small @xmath145 s contribute significantly , the lower limit of the integral over @xmath137 may be approximated simply by @xmath35 , so only a small `` initial '' error might be introduced . with these approximations equation yields @xmath160 ~+~ \mathbb{h.c}. \label{me72 } \end{aligned}\ ] ] introducing the quantities [ me73 ] @xmath161 , \label{me73a } \\ w_{\alpha\beta}(\omega ) & = \int_{0}^{\infty } d\tau ~e^{i \omega \tau } ~\bar{g}_{\alpha \beta}(\tau ) = \int_{0}^{\infty } d\tau ~e^{i \omega \tau } \ : { \mathrm{tr}}_{b } \bigl\ { \widetilde{x}^{\dagger}_{\alpha}(\tau ) \ , x_{\beta } { \bar{\sigma}_{b}}\bigr\ } , \label{me73b } \ ] ] we rewrite eq . as follows @xmath162 ~+~ \mathbb{h.c}. \label{me74 } \end{aligned}\ ] ] this equation specifies the rate of change of the reduced density operator @xmath42 within the time interval @xmath163 . the quotient @xmath164 can be treated as an averaging @xmath165 this averaging results in smoothing all very rapid changes of @xmath42 which may occur during the interval @xmath166 . in principle we should account for such rapid changes . we do not do that because right hand side of eq . contains @xmath42 , while the left hand side represents the smoothed rate of change . this rate depends on the density operator @xmath167 in the past , that is at the moment when the smoothed evolution was started . so our next approximation consists in replacing the smoothed rate by a usual derivative . in other words , the variation at an instant @xmath35 ( that is the derivative @xmath168 ) is connected with the value of @xmath42 at the very same instant . this approximation allows us to use a usual derivative at the left hand side of . this approximation sometimes is called a markovian one since it connects the variations of some physical quantity with its value at the same instant , independently from the values which this quantity had at earlier moments . we can say that markovian approximation consists in neglecting the influence of the history of the physical system on its current state which fully determines the presently occurring changes . in some literature sources this approximation is also called the coarse - graining one , because small and rapid fluctuations are neglected when the evolution is investigated on a much longer time scale specified by @xmath60 . with all the discussed approximation our master equation becomes @xmath169 ~+~ \mathbb{h.c}. \label{me76 } \end{aligned}\ ] ] at this stage , we return to the schrdinger picture and we insert @xmath170 into eq . . when computing the derivative at the left hand side we reproduce the free evolution term thus , from we get @xmath171 e^{-ih_{a}t/\hbar } \nonumber \\ & \hspace{1 mm } + ~\biggr\ { \ ; \frac{1}{\hbar^{2 } } \sum_{\omega,\omega\ , ' } \sum_{\alpha,\beta } j(\omega\,'- \omega ) w_{\alpha\beta}(\omega ) \bigl [ \ : a_{\beta}(\omega ) \ : e^{ih_{a}t/\hbar}\:\rho_{a}(t)\ : e^{-ih_{a}t/\hbar } a^{\dagger}_{\alpha}(\omega\ , ' ) \nonumber \\ & \hspace{29 mm } -~ a^{\dagger}_{\alpha}(\omega\ , ' ) a_{\beta}(\omega ) e^{ih_{a}t/\hbar } \ : \rho_{a}(t ) \ : e^{-ih_{a}t/\hbar } \bigr ] ~+~ \mathbb{h.c } \ ; \biggr\}. \label{me78 } \end{aligned}\ ] ] multiplying on the left by @xmath172 and on the right by @xmath173 , we use relation and its hermitian conjugate ( for negative times ) . this yields @xmath174 + ~\biggr\ { \ ; \frac{1}{\hbar^{2 } } \sum_{\omega,\omega\ , ' } \sum_{\alpha,\beta } j(\omega\,'- \omega ) w_{\alpha\beta}(\omega ) ~e^{i(\omega - \omega\,')t } \nonumber \\ & \hspace{12 mm } \times \bigl [ \ : a_{\beta}(\omega ) \rho_{a}(t ) a^{\dagger}_{\alpha}(\omega\ , ' ) ~-~a^{\dagger}_{\alpha}(\omega\ , ' ) a_{\beta}(\omega ) \rho_{a}(t ) \ : \bigr ] ~+~ \mathbb{h.c } \ ; \biggr\}. \label{me81 } \end{aligned}\ ] ] our master equation contains the integral @xmath175 defined in . its computation is straightforward . denoting temporarily @xmath176 , we get @xmath177 where we have introduced ( as in @xcite ) a function specified as @xmath178 due to the obtained results we can write @xmath179 inserting the computed integral into we note that the exponential factor cancels out . hence @xmath180 + \biggr\ { \ : \frac{1}{\hbar^{2 } } \sum_{\omega\,',\omega } \sum_{\alpha,\beta } f(\omega\,'-\omega ) w_{\alpha\beta}(\omega ) \nonumber \\ & \hspace{12 mm } \times \bigl [ \ : a_{\beta}(\omega ) \ : \rho_{a}(t ) \ : a^{\dagger}_{\alpha}(\omega ) ~-~a^{\dagger}_{\alpha}(\omega ) \ : a_{\beta}(\omega ) \ : \rho_{a}(t ) \ : \bigr ] ~+~ \mathbb{h.c } \ : \biggr\}. \label{me88}\end{aligned}\ ] ] the sense of function @xmath181 must be carefully considered . we proceed along the lines similar to those in @xcite . it is easy to see that function @xmath182 has a sharp maximum for @xmath183 , where it is equal to unity . [ 0.6 ] zeroes of this function correspond to @xmath184 if the time @xmath60 is sufficiently long then the central maximum is very narrow . the question is what does it mean `` sufficiently long time '' . let us consider two possibilities . 1 . if @xmath185 , the argument of function @xmath182 is very close to zero , its value being practically one . 2 . if @xmath186 ( bohr frequencies are significantly different ) then @xmath182 is close to zero , as it is seen in fig [ xmerys03 ] . we conclude that the terms at the right hand side of master equation containing the operator products @xmath187 , for which @xmath188 practically do not contribute to the evolution of the density operator @xmath0 . according to the first possibility above , significant contributions come only from such couplings that operators @xmath189 and @xmath190 have practically equal bohr frequencies . as we know , time @xmath45 is a characteristic relaxation time in subsystem @xmath18 due to interaction with reservoir and it satisfies the estimate @xmath191 ( we discuss it later ) . it can be argued ( see also @xcite ) that the terms in master equation , in which @xmath192 also give very small contributions , so that they can be neglected . as a result of all these approximations , we may say that only those terms in right hand side of master equation contribute significantly for which @xmath193 . such an approximation is called the secular one . it allows us to replace the function @xmath194 by the kronecker delta @xmath195 defined as in . it reminds us that only the terms satisfying the requirement @xmath196 give nonzero contribution . all these arguments lead to master equation of the form @xmath197 ~+~\biggr\ { \ : \frac{1}{\hbar^{2 } } \sum_{\alpha,\beta } \sum_{\omega\,',\omega } \delta(\omega\,'-\omega ) \ : w_{\alpha\beta}(\omega ) \nonumber \\ & \hspace{12 mm } \times \bigl [ \ : a_{\beta}(\omega ) \ : \rho_{a}(t ) \ : a^{\dagger}_{\alpha}(\omega\ , ' ) ~-~a^{\dagger}_{\alpha}(\omega\ , ' ) \ : a_{\beta}(\omega ) \ : \rho_{a}(t ) \ : \bigr ] ~+~ \mathbb{h.c } \ : \biggr\}. \label{me90 } \end{aligned}\ ] ] the presence of the discussed kronecker delta simplifies one of the summations , which gives @xmath198 + ~\biggr\ { \ : \frac{1}{\hbar^{2 } } \sum_{\alpha,\beta } \sum_{\omega } w_{\alpha\beta}(\omega ) \bigl [ a_{\beta}(\omega ) \rho_{a}(t ) a^{\dagger}_{\alpha}(\omega ) \nonumber \\ & \hspace{55 mm } ~-~ a^{\dagger}_{\alpha}(\omega ) a_{\beta}(\omega ) \rho_{a}(t ) \bigr ] + \mathbb{h.c } \ ; \biggr\}. \label{me91 } \end{aligned}\ ] ] the fundamental part of the microscopic derivation of the master equation is finished . we shall perform some transformations which have important , but rather cosmetic character . we want to transform master equation into the so - called standard form . all other comments are , as mentioned many times , left to _ auxiliary _ sections . standard form is important , because it can be shown ( see @xcite ) that this form guarantees preservation of hermiticity , normalization and , first of all , the positivity of the reduced density operator @xmath47 . if our master equation can be brought into the standard form , then we can be sure that all the necessary properties of the density operator of subsystem @xmath1 are indeed preserved . obviously , the first term in the right hand side of equation describes the unitary evolution , hence we shall concentrate only on the second term . writing explicitly the hermitian conjugates , we have @xmath199 \nonumber \\ & + \frac{1}{\hbar^{2 } } \sum_{\omega } \sum_{\alpha,\beta } w^{\ast}_{\alpha\beta}(\omega ) \bigl [ \ : a_{\alpha}(\omega ) \rho_{a}(t ) a^{\dagger}_{\beta}(\omega ) ~-~\rho_{a}(t ) a^{\dagger}_{\beta}(\omega ) a_{\alpha}(\omega ) \ : \bigr ] , \label{me96a } \end{aligned}\ ] ] because operator @xmath0 is hermitian ( the proof that hermiticity is preserved will be presented in _ auxiliary _ sections ) . in the second term we interchange the summation indices @xmath200 which gives @xmath199 \nonumber \\ & \hspace{4 mm } + \frac{1}{\hbar^{2 } } \sum_{\omega } \sum_{\alpha,\beta } w^{\ast}_{\beta \alpha}(\omega ) \bigl [ \ : a_{\beta}(\omega ) \rho_{a}(t ) a^{\dagger}_{\alpha}(\omega ) ~-~\rho_{a}(t ) a^{\dagger}_{\alpha}(\omega ) a_{\beta}(\omega ) \ : \bigr ] . \label{me96b } \end{aligned}\ ] ] for further convenience we introduce the following notation [ me97 ] @xmath201 . \label{me97b } \ ] ] the matrix @xmath202 is hermitian and positively defined . the latter property is difficult to prove . it requires some advanced mathematics and we take this fact for granted . the readers are referred to literature @xcite . hermiticity of @xmath203 follows directly from the definition . indeed , we have @xmath204 the second matrix @xmath205 is also hermitian . from it follows that @xmath206 = \frac{1}{2i } \bigl [ w_{\beta\alpha}(\omega ) - w^{\ast}_{\alpha \beta}(\omega ) \bigr ] = \delta_{\beta\alpha}(\omega ) . \label{me100}\ ] ] let us focus on the method of computation of elements @xmath202 . as it will be shown , elements @xmath205 are less important . to find @xmath207 we need quantities @xmath208 . conjugating definition we find that @xmath209 where we used relations , and cyclic property of trace . changing the integration variable @xmath210 , we have @xmath211 the integrand is identical as in the definition , only the integration limits are different . combining this with , we get @xmath212 the elements @xmath202 are the fourier transforms of the corresponding correlation function of the reservoir . matrix @xmath205 does not have such a simple representation . from the definition and the second relation in @xmath213 . \label{me101h } \end{aligned}\ ] ] inverting relations we express elements @xmath214 via @xmath207 and @xmath215 . after simple regrouping of the terms in eq . we find @xmath216_{+ } - i \delta_{\alpha \beta}(\omega ) \bigl [ a^{\dagger}_{\alpha}(\omega ) a_{\beta}(\omega ) , \ : \rho_{a}(t ) \bigr ] \ : \bigr\}. \label{me102d } \end{aligned}\ ] ] let us note that the last term is a commutator , so we define @xmath217 taking into account hermiticity of matrix @xmath218 ( changing the names of the summation indices when necessary ) we can easily show that the operator @xmath219 is also hermitian . returning to full master equation , that is to eq . , we conclude that the term containing @xmath219 in can be connected with the free hamiltonian one . in this manner , we finally have @xmath220 \nonumber \\ & + ~\frac{1}{\hbar^{2 } } \sum_{\omega } \sum_{\alpha,\beta } \ ; \gamma_{\alpha \beta}(\omega ) \ : \bigl\ { \ : a_{\beta}(\omega ) \ : \rho_{a}(t ) \ : a^{\dagger}_{\alpha}(\omega ) \nonumber \\ & \hspace{40 mm } - { { \textstyle \frac{1}{2}}}\ : \bigl [ a^{\dagger}_{\alpha}(\omega ) \ : a_{\beta}(\omega ) , \ : \rho_{a}(t ) \bigr]_{+ } \bigr\ } , \label{me106 } \end{aligned}\ ] ] which coincides exactly with the standard form of the evolution equation for the reduced density operator @xmath0 which describes the state of the subsystem @xmath1 interacting with reservoir @xmath2 . this allows us to be sure that hermiticity , normalization and positivity of the operator @xmath0 are indeed ensured . finally , let us remark that operator @xmath219 which gives a contribution to the hamiltonian ( unitary ) evolution , usually produces small shifts of the eigenenergies of the subsystem @xmath1 . that is why , in many practical applications , this term is simply omitted . this explains our previous remark that matrix @xmath221 is less important than @xmath203 . obviously one can construct operator @xmath219 and investigate its influence on the unperturbed energy levels of the subsystem @xmath1 . small energy shifts of eigenenergies of subsystem @xmath1 are qualitatively similar to the well - known lamb shifts , which clarifies the employed notation . the obtained master equation is an operator one . in practice , we frequently need an equation of motion for the matrix elements of the reduced density operator @xmath0 . it seems to be natural to use the energy representation , that is to consider matrix elements of @xmath0 calculated in the basis @xmath222 of the eigenstates of the free hamiltonian @xmath16 ( see eq . ) . this will be done in the next section . when analyzing master equation in the basis of the eigenstates of free hamiltonian we need to be careful . the reason is that the commutator in contains an additional term , namely the lamb - shift hamiltonian . one may argue that this changes the hamiltonian and a new basis should be found ( a basis in which @xmath223 is diagonal ) . we will , however , proceed in the spirit of the perturbative approach . we will treat @xmath219 as a small perturbation which , at most , will yield small energy shifts . therefore , the set @xmath224 of eigenstates of the unperturbed hamiltonian @xmath16 can be used as complete and orthonormal basis . working within this scheme , we can easily construct master equation ( equation of motion ) for matrix elements of the density operator for subsystem @xmath1 . we will suppress the index @xmath225 since it should lead to no misunderstanding . taking matrix elements @xmath226 and expanding the anticommutator term we obtain @xmath227 { \| \ , b \ , \rangle } \nonumber \\ & + ~ \frac{1}{\hbar^{2 } } \sum_{\omega } \sum_{\alpha,\beta } \ : \gamma_{\alpha \beta}(\omega ) \ : \bigl\ { \ : { \langle \ , a \ , \| } a_{\beta}(\omega ) \ : \rho(t ) \ : a^{\dagger}_{\alpha}(\omega ) { \| \ , b \ , \rangle } \nonumber \\ & \hspace{-5 mm } - { { \textstyle \frac{1}{2}}}\ : { \langle \ , a \ , \| } a^{\dagger}_{\alpha}(\omega ) \ : a_{\beta}(\omega ) \ : \rho(t ) { \| \ , b \ , \rangle } - { { \textstyle \frac{1}{2}}}\ : { \langle \ , a \ , \| } \rho(t ) \ : a^{\dagger}_{\alpha}(\omega ) \ : a_{\beta}(\omega ) { \| \ , b \ , \rangle } \bigr\}. \label{me112 } \end{aligned}\ ] ] the last three terms constitute a so - called dissipative term ( or dissipator ) and we will concentrate on its form . first , we use expressions , for operators @xmath228 and @xmath100 . then we consider three matrix elements . necessary computations in the basis of eigenstates of free hamiltonian @xmath16 are simple though a bit tedious , in some cases a suitable changes of summation indices is necessary . the results of these calculations are as follows [ me117 ] @xmath229 & \hspace{9 mm } = ~\sum_{m , n } \delta(\omega_{ma } - \omega ) \ : \delta(\omega_{nb } - \omega ) { { \langle \ , a \ , \|}\,a_{\beta}\,{\| \ , m \ , \rangle } } { { \langle \ , n \ , \|}\,a^{\dagger}_{\alpha}\,{\| \ , b \ , \rangle } } \ : \rho_{mn}(t ) , \label{me117a } \end{aligned}\ ] ] @xmath230 & \hspace{9 mm } = \sum_{m , n } \delta(\omega_{an } - \omega ) \ : \delta(\omega_{mn } - \omega ) { { \langle \ , a \ , \|}\,a^{\dagger}_{\alpha}\,{\| \ , n \ , \rangle } } { { \langle \ , n \ , \|}\,a_{\beta}\,{\| \ , m \ , \rangle } } \ : \rho_{mb}(t ) , \label{me117b } \end{aligned}\ ] ] @xmath231 & \hspace{9 mm } = \sum_{m , n } \delta(\omega_{mn } - \omega ) \ : \delta(\omega_{bn } - \omega ) { { \langle \ , m \ , \|}\,a^{\dagger}_{\alpha}\,{\| \ , n \ , \rangle } } { { \langle \ , n \ , \|}\,a_{\beta}\,{\| \ , b \ , \rangle } } \ : \rho_{am}(t ) . \label{me117c } \ ] ] the computed matrix elements are plugged into equation and summation over frequency @xmath103 is performed . after some regrouping we find that @xmath232 & \hspace{13 mm } -~{{\textstyle \frac{1}{2}}}\ : \gamma_{\alpha \beta}(\omega_{mn } ) \ : \delta(\omega_{bn } - \omega_{mn } ) { { \langle \ , n \ , \|}\,a_{\beta}\,{\| \ , b \ , \rangle } } { { \langle \ , n \ , \|}\,a_{\alpha}\,{\| \ , m \ , \rangle}}^{\ast } \ : \rho_{am}(t ) \bigr\}. \label{me118b } \end{aligned}\ ] ] going further , we use the evenness of kronecker delta in the first term , while the presence of the deltas in the second and third terms allows us to change arguments in the elements of matrix @xmath203 . next , we denote @xmath233 due to these , we rewrite formula as @xmath234 & \hspace{40 mm } -~{{\textstyle \frac{1}{2}}}\ : \sum_{m , n } \ : \delta(\omega_{bn } - \omega_{mn } ) \ ; k(nb , nm ) \ ; \rho_{am}(t ) . \label{me120 } \end{aligned}\ ] ] let us note the specific symmetry of this expression . further analysis depends on whether the eigenfrequencies of the hamiltonian @xmath16 are degenerate or not . we also note that kronecker deltas in the second and third terms are correspondingly given as @xmath235 and @xmath236 , which allows one to perform summation over @xmath237 . however , one has to be careful because eigenfrequencies @xmath238 can be degenerate . to account for the possible degeneracies , let us write the hamiltonian of the considered system @xmath1 in the following form @xmath239 where @xmath240 is the main quantum number which distinguishes energy levels ( energy multiplets ) , while @xmath241 , are subsidiary quantum numbers . is is obvious that @xmath242 for @xmath243 . certainly , the nondegenerate case follows immediately and it corresponds to @xmath244 , then subsidiary quantum numbers are unnecessary and can be simply suppressed . in the degenerate case single indices appearing in equation must be replaced by corresponding pairs , for example @xmath245 . equation is now rewritten as @xmath246 & \hspace{10 mm } -~{{\textstyle \frac{1}{2}}}\ : \sum_{mm } \sum_{nn } \ : \delta(\omega_{bn } - \omega_{mn } ) \ ; k(nnbb , nnmm ) \ ; \rho_{aamm}(t ) . \label{me131 } \end{aligned}\ ] ] as already noted , one immediately sees that @xmath247 and similarly @xmath248 , where the last deltas are the simple kronecker ones . the sum over @xmath249 in the second term is trivial . we put @xmath250 and we `` land within multiplet @xmath225 '' , hence we change @xmath251 . analogously , in the second term @xmath252 and @xmath253 . therefore , we have @xmath254 & \hspace{40 mm } -~{{\textstyle \frac{1}{2}}}\ : \sum_{nn } \sum_{b '' } k(nnbb , nnbb '' ) \ ; \rho_{aabb''}(t ) . \label{me132 } \end{aligned}\ ] ] in two last terms matrix elements do not depend on quantum numbers @xmath255 , hence we can denote @xmath256 this allows us to write equation in the form @xmath257 let us consider this equation in some more detail . first , we take @xmath258 ( and correspondingly @xmath259 ) . this yields the equation of motion for `` quasi - population '' matrix elements taken within just one energy multiplet . then , the first term in right - hand side contains @xmath260 . the sum over @xmath261 is trivial ( @xmath262 ) and we have @xmath263 this equation connects `` quasi - populations '' with other ones . the first sum contains the term with @xmath264 and this term represents elastic ( energy conserving ) processes . the remaining terms ( with @xmath265 ) corresponding to nonelastic transitions . in this case , the environment @xmath2 serves as a reservoir which gives or absorbs the energy . the terms in the second line describe the `` escape '' from multiplet @xmath225 to other ones . to discuss coherences , we assume @xmath266 , which implies @xmath267 . the kronecker delta in can be rewritten as @xmath268 . since @xmath269 , we also get @xmath270 . if we assume that all energy distances are different ( that is @xmath271 for different pairs @xmath272 ) the considered delta can give unity only when @xmath264 and @xmath273 ( which entails @xmath274 and @xmath275 then , eq . reduces to @xmath276 so the coherences between two multiplets @xmath266 couple only with coherences from just these multiplets . obviously for the nondegenerate case `` small '' indices play no role they can be suppressed . then , instead of equation for `` quasi - populations '' we get an equation for genuine populations @xmath277 similarly for coherences , eq . yields @xmath278 \;\rho_{ab}(t ) . \label{me148}\ ] ] these examples indicate that me for matrix elements of the reduced density operator possess quite a specific symmetry which probably can be further investigated . this , however , goes beyond the scope of the present work . any density operator , so also the reduced one for subsystem @xmath18 must be normalized , that is , we require that @xmath279 . this has a simple consequence @xmath280 clearly the hamiltonian part ( the commutator ) preserves the trace , which follows from cyclic property . hence we must check the second dissipative part of our me . one may ask at which stage of our derivation such a check should be made . in principle , this can be done at any stage . in this section we shall do so twice . once for standard form , and for me in the energy basis . taking me in its standard form we need to compute the following trace @xmath281 \bigr\ } , \label{mea03 } \end{aligned}\ ] ] and show that it vanishes , ie . , @xmath282 . the trace is a linear operation , so then @xmath283 & \hspace{42 mm } ~-~ { { \textstyle \frac{1}{2}}}\ : \gamma_{\alpha \beta}(\omega ) \ : { \mathrm{tr}}_{a } \bigl\ { \rho_{a}(t ) \ : a^{\dagger}_{\alpha}(\omega ) \ : a_{\beta}(\omega ) \bigr\ } \bigr ] . \label{mea04 } \end{aligned}\ ] ] cyclic property allows one to see that all three traces are equal . therefore , @xmath282 and we conclude that preservation of the normalization for me in the standard form is proved . in this case we check the trace preservation for eq . , with @xmath284 . we need to compute @xmath285 in the first term we use definition of the parameter @xmath286 ( see ) . in the two next ones we notice that indices @xmath93 and @xmath287 concern the same multiplet @xmath225 , so the summation range is also the same . we can interchange @xmath288 and obtain @xmath289 the second and third terms are identical and cancel out with the first one ( names of summation indices are irrelevant ) . we have shown that in the energetic basis the trace of the reduced density operator for subsystem @xmath1 is preserved . in other words , the derived me preserves normalization . the next necessary property of any density operator is its hermiticity . if the equation of motion for @xmath290 is identical with the similar equation for @xmath0 , then the same equations must yield the same solutions , this means that @xmath291 . free evolution is given by the hamiltonian term @xmath292 $ ] which poses no problems due to the commutator properties . one needs to investigate the dissipative part of me . as in the previous section we perform such a check for me in standard form and for the one in energy basis . we take the hermitian conjugate of the dissipative part of me @xmath293_{+ } \bigr\ } , \label{meb01 } \end{aligned}\ ] ] because conjugate anticommutator is equal to the anticommutator of conjugated operators . we know ( see ) that matrix @xmath203 is hermitian . interchanging the indices @xmath200 we get @xmath294_{+ } \bigr\ } , \label{meb02 } \end{aligned}\ ] ] we see that the equation of motion for @xmath295 is formally identical with standard form of me . thus , @xmath296 hermiticity is preserved . hermiticity of the density operator means that @xmath297 . it entails , that the equation of motion for the element @xmath298 must be the same as for @xmath299 . starting from me we look for a corresponding equation for @xmath298 . first we need to change the indices in ( remembering that corresponding changes must be made for summation indices in all terms ) and then we perform complex conjugation . in this manner we find @xmath300 next , we need to consider the conjugated quantities @xmath301 and @xmath302 . by definition @xmath303 we take complex conjugates , use hermiticity of matrix @xmath207 and we notice that the presence of the kronecker delta allows the change of the argument in @xmath304 . interchanging the summation indices @xmath200 we have @xmath305 comparing this relation with definition we see that @xmath306 next , we deal with the parameter @xmath307 . in the above relation we substitute @xmath308 , @xmath309 , @xmath310 and @xmath311 . then @xmath312 obviously kronecker deltas are equal to one , so they are unimportant . using this result in the definition of the parametr @xmath286 we get @xmath313 returning to the analysis of formula , we use the proven relations and . at the same time , in the first term in the right hand side we interchange the summation indices @xmath314 . moreover , we recall that kronecker delta is even . thus , we have @xmath315 comparing this result with eq . we find that when we replace @xmath299 by @xmath298 ( consequently in all the terms ) then we will arrive at . this equation is formally identical with eq . , hence @xmath316 , what we intended to show . our me preserves hermiticity of the reduced density operator of subsystem @xmath1 . in this section we will show that the reduction of the density operator is invariant with respect to the choice of the picture . the definition of the reduced density operator and for the transformation to the interaction picture imply that the operator @xmath0 ( in the schrdinger picture ) is expressed as @xmath317 we note that the free evolution of each of the subsystems written as @xmath318 can be expressed with the aid of the operator @xmath319 because both hamiltonians are fully independent and commute . in eq . we compute the trace only over reservoir variables , so we can write @xmath320 cyclic property of the trace yields @xmath321 left hand side represents the reduced density operator in the interaction picture ( it depends solely on the variables of the subsystem @xmath18 ) . hence , we have @xmath322 this is formally identical with the definition of the reduced density operator in the schrdinger picture . the relation between the reduced density and the total one is the same in both pictures . in other words , reduction of the operatora @xmath323 is invariant with respect to the change of the pictures . in this section we follow and ( hopefully ) try to to elucidate the discussion presented by cohen - tannoudji _ et al _ @xcite . the previously given remarks apply also here , so we again refrain from frequent citations . the key role in our considerations is played by the assumption . this is the requirement @xmath324 in other words we assume that there exist two , quite distinct , time scales . firstly , let us try to estimate the time @xmath45 which characterizes the evolution of system @xmath18 due to the interaction with reservoir . to find such an estimate we use eq . , that is @xmath325 \right ] \biggl\ } , \label{med02 } \end{aligned}\ ] ] where we employed the introduced properties of the reservoir . we also recall that the main contribution to the integrals comes from a thin belt ( of width @xmath44 lying below the diagonal @xmath326 , see fig . [ xmerys01 ] and its discussion ) . this allows us to estimate the integrand as follows @xmath327 \right ] ~\sim~ { \widetilde{\varrho}}_{a } \ : { \mathrm{tr}}_{b } \left\ { \ : \widetilde{v}^{2 } { \bar{\sigma}_{b}}\ : \right\ } ~=~ v^{2 } { \widetilde{\varrho}}_{a}. \label{med03}\ ] ] hence , left hand side of eq . is estimated by @xmath328 because the area of the integration region is estimated by the product @xmath329 ( area of the belt under the diagonal @xmath140 ) . introduced parameter @xmath50 characterizes the `` strength '' of the interaction between the reservoir and system @xmath18 . the factor which multiplies @xmath167 in has ( according to ) the sense of the inverse of time @xmath45 . therefore , we obtain an estimate @xmath330 what is the condition of the existence of two time scales ? the estimate of @xmath45 given in allows us to find such a condition . let us look upon condition @xmath52 more carefully and introduce the estimate . this yields @xmath331 the last inequality is the sought condition of existence of two time scales . if we denote @xmath332 , then we can write @xmath333 . so the interaction must be characterized by such bohr frequency @xmath334 that during the time interval of magnitude of @xmath44 its influence on system @xmath18 is negligibly small . we already mentioned ( see the discussion of eq . ) , that it is possible to iterate von neumann equation accounting for higher order corrections would increase accuracy . we can estimate these higher order terms in the same manner as done just above . for example , for the third order term we have @xmath335 because times @xmath136 , @xmath143 and @xmath336 must be close to each other ( with accuracy of the order of @xmath44 ) . then the region of integration has volume of the order of @xmath337 . due to eq . we get @xmath338 where the last inequality follows from . the obtained estimate shows that the third order iteration ( and similarly higher ones ) are indeed negligible . obviously this holds provided the condition is true and ensures the existence of two distinct time scales . moving from eq . to we have neglected initial correlations between systems @xmath18 and @xmath19 . these correlations built up at earlier moments @xmath340 . this corresponds to the assumption that at some earlier moment @xmath341 ( @xmath342 ) both systems were uncorrelated . this happens , for example , when the interaction was switched on at an instant @xmath341 . so the correlations described by @xmath343 need time @xmath344 to appear . if the correlations exist ( @xmath345 ) then averaging of the term linear in interaction ( as in expression ) would not give zero . hence , @xmath345 would result in the appearance of the linear term . moreover , this would also automatically modify the quadratic term in . let us estimate the magnitude of this modification ( which is due to earlier interaction ) @xmath346 the integrand contains correlation functions of the reservoir . hence the integrand would be practically zero for @xmath347 . the integration runs effectively from @xmath348 to @xmath349 . therefore , using condition , we estimate @xmath350 as the integrals are nonzero within the interval of the length of the order of @xmath44 . if @xmath61 then the correction is small ( main contribution to the evolution of @xmath167 is of the order of @xmath351 , which is quite larger ) . the key assumption that @xmath61 allows us to conclude that the correlations between system @xmath18 and @xmath19 which were built before moment @xmath35 do not significantly change the evolution of @xmath42 , their influence is restricted to the moments from a very short interval @xmath352 . new correlations , within a much longer interval @xmath163 , are building up and have an effect on the evolution of @xmath42 . initial correlations have small significance and hence it is justified to neglect them . secular approximation consists in replacing the function @xmath181 ( defined in ) in by kronecker delta , which leads to eq .. our discussion of this replacement does not rise any doubts when @xmath353 , because then @xmath194 is practically unity . on the other hand for @xmath354 the function @xmath194 is practically zero . the only problem is to justify the neglecting of the terms for which @xmath355 . to explain this point , first use the fact that the free evolution of matrix elements @xmath356 is governed by @xmath357 { \| \ , b \ , \rangle } ~=~ -i \omega_{ab } \rho_{ab}^{(f ) } , \label{medd01}\ ] ] where the small lamb shift ( due to @xmath219 ) is ignored . the solution is simple @xmath358 now , we analyze the dissipative term which is given by . we do not discuss the nuances connected with possible degeneracies . obviously , we can write @xmath359 which we use in . we also interchange indices @xmath360 in the second term and similarly , in the third one we first interchange @xmath361 and then @xmath362 . the result is @xmath363 we note that @xmath364 implies @xmath365 , and then @xmath366 . therefore @xmath367 since @xmath366 and changes nothing . similarly we have @xmath368 finally , we note that @xmath369 , because kronecker delta is even . after these manipulations eq . can be written as @xmath370 the expression in braces is denoted as @xmath371 and we have @xmath372 this is a specific form of me , but useful in the discussion . however , in the degenerate case some care must be exercised and renewed considerations might be necessary . we proceed to the discussion of conditions and/or possibilities of neglecting the terms for which @xmath355 . interaction with the reservoir certainly modifies the free evolution of @xmath373 . if bohr frequencies of the oscillating elements @xmath356 i @xmath374 are such that @xmath375 , then mutual couplings between these elements are quickly averaged to zero ( interfere destructively ) before time @xmath45 elapses and before the interaction has enough time to affect the evolution . in other words , if @xmath376 differs much from @xmath351 then the coupling between corresponding matrix elements will have small ( weak ) effect . this is the situation similar to the one encountered in perturbation theory . namely , when the energies @xmath377 then the perturbation has small ( usually negligible ) effect . since , by assumption @xmath378 the discussed situation corresponds , in fact , to the relation @xmath379 . this , in turn means , that such terms have little influence on the evolution of the operator @xmath0 . such terms are neglected while passing from eq . thus the last of our approximations is justified . in our considerations we have adopted the interaction hamiltonian between the system @xmath18 and reservoir @xmath19 in the form @xmath381 , where operators @xmath75 and @xmath76 do not have to be hermitian . certainly , the full interaction hamiltonian must be hermitian , so we conclude that it must contain operators @xmath75 , @xmath76 and their hermitian conjugates @xmath382 , @xmath383 . constructing linear combinations we can always transform the interaction hamiltonian into @xmath384 , where the primed operators are hermitian . we shall illustrate this with a simple example . let the interaction hamiltonian be of the form @xmath385 where operators @xmath225 and @xmath386 are nonhermitian , while the full hamiltonian is clearly hermitian . we define new operators @xmath387 which are evidently hermitian . expressing operators @xmath225 , @xmath386 and their conjugates via the new ones , we obtain @xmath388 this interaction hamiltonian is expressed as a sum of products of hermitian operators . hence construction of the interaction hamiltonian with nonhermitian operators is allowed . one can always build necessary combinations . however , in some practical applications it is much more convenient to use nonhermitian operators than the linear combinations . in the main part of the lecture we assumed that eq . holds , that is the average @xmath390 . we have stated that it is not really restrictive . we will show that it is true . this is so , because we can always shift the energy scale . to see this , let us write @xmath391 where @xmath392 is a number not necessarily equal to zero . then we have @xmath393 which holds no matter whether numbers @xmath389 are zeroes or not . full hamiltonian can then be written as @xmath394 \otimes { \mathbf{1}}_{b } ~+~{\mathbf{1}}_{a}\otimes h_{b } ~+~v^{\,'}_{ab}. \label{mef3 } \end{aligned}\ ] ] rescaled interaction term ( the last one ) has zero average ( as in ) . this is achieved by the redefinition of the energy scale in system @xmath18 via redefinition of the hamiltonian @xmath16 . we conclude that the assumption that the averages vanish is not really restrictive , but simplifies the computation . in the main sections we have introduced the operators @xmath87 defined by relation . the hamiltonian of system @xmath1 is of the form @xmath395 . it is not difficult to find the commutator @xmath396 $ ] . directly from the definitions we obtain @xmath397 = \bigr[\sum_{n } \hbar\omega_{n } { \| \ , n \ , \rangle}{\langle \ , n \ , \| } , ~\sum_{a , b } \delta(\omega_{ba } - \omega ) \ : { \| \ , a \ , \rangle}{{\langle \ , a \ , \|}\,a_{\alpha}\,{\| \ , b \ , \rangle}}{\langle \ , b \ , \| } \bigl ] \nonumber \\ & \hspace{10 mm } = \sum_{a , b } \hbar ( \omega_{a } - \omega_{b } ) \ : \delta(\omega_{ba } - \omega ) { \| \ , a \ , \rangle } { { \langle \ , a \ , \|}\,a_{\alpha}\,{\| \ , b \ , \rangle}}{\langle \ , b \ , \| } \nonumber \\ & \hspace{10 mm } = - \ ; \hbar \omega \sum_{a , b } \delta(\omega_{ba } - \omega ) { \| \ , a \ , \rangle } { { \langle \ , a \ , \|}\,a_{\alpha}\,{\| \ , b \ , \rangle}}{\langle \ , b \ , \| } ~=~ - \hbar \omega a_{\alpha}(\omega ) , \label{meg03 } \end{aligned}\ ] ] which ends the calculation . conjugation changes sign , so that @xmath398 ~=~ \hbar \omega a_{\alpha}^{\dagger}(\omega ) . \label{meg04}\ ] ] heisenberg equation of motion follows from formula , and it is @xmath399 ~=~ \hbar \omega a_{\alpha}^{(h)}(\omega ) . \label{meg05}\ ] ] after integration we obtain @xmath400 which agrees with . finally , we present one more relation @xmath401 = a^{\dagger}_{\alpha}(\omega ) \bigl [ \,h_{a } , \ ; a_{\beta}(\omega ) \ : \bigr ] \nonumber \\ & \hspace{50 mm } ~+~ \bigl [ \,h_{a } , \ ; a_{\alpha}^{\dagger}(\omega ) \ : \bigr ] a_{\beta}(\omega ) = 0 , \label{meg09 } \end{aligned}\ ] ] which follows immediately from the derived results . correlation function of the reservoir was defined in or . by assumption , reservoir hamiltonian @xmath17 and the corresponding density operator @xmath70 commute , so they have a common set of complete and orthonormal eigenstates @xmath66 . let us calculate the trace in in the chosen basis @xmath403 in eq . we denoted the eigenvalues of @xmath70 by @xmath404 , hence @xmath405 with @xmath406 , and @xmath407 . expression shows that the correlation function @xmath402 is a complicated superposition of functions which oscillate with bohr frequencies @xmath408 . reservoir is assumed to be large , the discussed frequencies are densely space ( quasi - continuous ) . if time @xmath145 is sufficiently large the oscillations interfere destructively ( average out to zero ) . we can expect that reservoir correlation function decay quickly when time @xmath123 increases . characteristic decay time is denoted by @xmath44 and assumed to be , by far , the shortest time characterizing the system @xmath14 . when @xmath409 the correlation may be neglected . in this summary we describe practical steps needed in the construction of the me for specified physical systems . the first step consists in precise definition of the system @xmath1 and of the reservoir @xmath2 . we need to specify their free hamiltonians @xmath16 and @xmath17 and ( at least sometimes ) their eigenenergies and eigenstates . then we define the interaction hamiltonian in the form @xmath410 where @xmath411 are ( correspondingly ) operators of system @xmath1 and reservoir . we stress that these operator need not be ( separately ) hermitian . it suffices that the full interaction hamiltonian is hermitian . we also need to specify the density operator @xmath70 describing the state of the reservoir . it is worth remembering that operator @xmath17 and @xmath70 commute . this implies that the reservoir is in the stationary state . in the second step of me construction we build ( identify ) the following operators @xmath412 the following matrix elements are computed in the third step @xmath413 they are seen to be partial fourier transform of the reservoir correlation functions . reservoir operators are taken in the interaction picture @xmath414 coefficients @xmath415 are then employed to construct two hermitian matrices @xmath416 . \label{mep05 } \end{aligned}\ ] ] we note that @xmath202 is a positive - definite matrix and can be computed directly as fourier transform @xmath417 parameters @xmath202 , in practical applications , are more important than @xmath205 . explanation will be given later . the separate expression for elements @xmath418 is @xmath419 . \label{mep6b}\ ] ] hence , calculation of coefficients @xmath415 can be usually omitted . final construction of the proper me is the fourth and the last step . the above given quantities allow us to write the me as @xmath420 \nonumber \\ & \hspace{8 mm } + ~\frac{1}{\hbar^{2 } } \sum_{\omega } \sum_{\alpha,\beta } \gamma_{\alpha \beta}(\omega ) \ : \big\ { \ : a_{\beta}(\omega ) \ : \rho_{a}(t ) \ : a^{\dagger}_{\alpha}(\omega ) \nonumber \\ & \hspace{45 mm } -~ { { \textstyle \frac{1}{2}}}\ : \bigl [ a^{\dagger}_{\alpha}(\omega ) \ : a_{\beta}(\omega ) , \ : \rho_{a}(t ) \bigr]_{+ } \bigr\ } , \label{mep07 } \end{aligned}\ ] ] where the so - called lamb - shift hamiltonian @xmath219 is given as @xmath421 energy shifts of the system @xmath1 which are due to the presence of @xmath219 in the hamiltonian part , are usually quite small and frequently negligible . this explains why the role of matrix @xmath221 is usually less important than that of matrix @xmath203 . 99 c. cohen - tannoudji , b. diu , f. lalo , _ quantum mechanics _ , + wiley - interscience , new york 1991 . r. alicki and k. lendi , _ quantum dynamical semigroups and applications _ , + lect . notes phys 717 , ( springer , berlin heidelberg 2007 ) . breuer , f. petruccione , _ the theory of open quantum systems _ , + oxford university press , 2002 . k. hornberger , _ introduction to decoherence theory _ , arxiv : quant - ph/0612118v2 . j. preskill , _ lecture notes on quantum computation _ , + http://www.theory.caltech.edu/@xmath422preskill/ph229 . c. cohen - tannoudji , j. dupont - roc , g. grynberg , _ atom photon interactions _ , + wiley , new york 1992 .	we do not present any original or new material . this is a tutorial addressed to students who need to study the microscopic derivation of the quantum - mechanical master equation encountered in many practical physical situations . = -1 mm
there has in the last few years been considerable interest in extended systems which self - organize into a state exhibiting large scale fluctuations and intermittent dynamics . one of the earliest attempts to model systems of this type was made in 1987 by bak , tang , and wiesenfeld , who proposed a simple lattice model for the avalanches produced by depositing grains of sand on an ever - growing sand pile @xcite . despite having only short - range interactions and no tunable parameters , their model organizes itself into a state with long - range spatial correlations and avalanches of size not limited by any finite correlation length . it has been proposed that similar self - organized critical ( soc ) behavior could lie behind a wide range of physical phenomena showing @xmath0 noise and scale - free fluctuation distributions . soc models have been put forward to describe the dynamics of earthquakes @xcite , biological evolution @xcite and extinction @xcite , forest fires @xcite , and many other systems @xcite . the common features of these models are that ( i ) they are all driven very slowly , and ( ii ) they all have perfect memory , i.e. , in the absence of the driving force the model would be entirely stationary . a distinctive observable consequence of soc dynamics is that the distribution of fluctuation ( or avalanche ) sizes takes a power - law form with characteristic exponent @xmath1 : @xmath2 the value of @xmath1 typically lies in the range @xmath3 , with the value @xmath4 corresponding to a critical branching process , appearing if one makes the `` random neighbor approximation '' in which each site interacts with a randomly - selected small number of other sites @xcite . this approximation is equivalent to the limit of infinite dimension , and should give correct results for systems above their critical dimension . in reality however , a number of the systems which are modeled using soc dynamics in fact display event size distributions with fairly large exponents . terrestrial earthquakes , for example , appear to follow the gutenberg richter law @xcite with @xmath5 , and the 1d rice pile experiment of frette and co - workers @xcite , which has been compared to the sandpile model of bak , tang and wiesenfeld , gives @xmath6 . a number of models have been proposed which offer explanations for these higher values of @xmath1 . the model discussed by christensen and olami @xcite is one such , but it achieves its result only at the price of having an entirely deterministic dynamics ; if one introduces randomness into the model , the simple scaling behavior is destroyed . the reason is that when the exponent describing the distribution of avalanches spatial extent becomes larger than @xmath7 , the mean avalanche size becomes finite and independent of system size , and the spatial overlap between subsequent avalanches becomes insignificant @xcite . in the presence of randomness , this can prevent the system from building up any long - range correlations , and ultimately destroy the critical state . we conjecture that , in this regime , any randomness in the positions of the nucleation centers of the avalanches will destroy self - organization of long range spatial correlations . in this paper , we present a different explanation to account for systems that have larger values of @xmath1 . we demonstrate that power - law event size distributions having @xmath1 around 2 or greater , are _ typical _ of extended systems with quenched memory if they are driven by coherent noise , and that in such systems they are present even in the absence of any interaction between the different parts of the system . ( this is different from the situation in the soc models , where the system is driven by a local driving force , coupled with interactions between the components of the system . ) the simplest model demonstrating the phenomenon is defined as follows . consider a system of @xmath8 agents , such as grains on the surface of a sand pile or points of contact in a subterranean fault . with each agent @xmath9 we associate a threshold for movement @xmath10 which can take values falling in some specified range and represents the amount of stress that the agent will withstand before it moves . for convenience , we choose to measure @xmath10 on a scale on which @xmath11 . the dynamics of the model then consists of the repetition of two steps : 1 . a fixed fraction @xmath12 of the agents are selected at random , and the values of their threshold variables @xmath10 are exchanged with new random numbers selected uniformly from the interval @xmath13 . we select a random number or `` stress level '' @xmath14 from some distribution @xmath15 . all @xmath10 below @xmath14 are also exchanged with new random numbers selected uniformly from the interval @xmath13 . the number of agents whose thresholds are changed in this fashion is the size @xmath16 of the avalanche taking place in this time - step . the random selection of different values for @xmath14 at each step may be thought of as imposing external stresses which _ coherently _ ( in other words , simultaneously ) influence all of the weaker agents those having suitably low thresholds for stress but leaves unchanged the stronger ones . it seems physically reasonable to assume that smaller stresses should be more common than larger ones , and in the following discussion we make the assumption that @xmath15 is largest at @xmath17 and falls off to zero as @xmath14 becomes large . we denote the typical scale of the falloff by @xmath18 . the most interesting regime is when @xmath19 and @xmath20 , we have examined the properties of this model both analytically and numerically . instead of simulating the model directly , we have developed an algorithm which calculates the threshold distribution and avalanche sizes in a formally exact way for a system with @xmath21 . starting off with a uniform distribution of thresholds , the system evolves towards a statistically stationary state . in this state we record the mean threshold distribution and the frequency distribution of avalanches . the results are shown in figures [ size ] and [ thresh ] for a simulation using exponentially distributed stresses : @xmath22 as figure [ size ] shows , the distribution of avalanche sizes @xmath16 is flat up to a certain point ( whose position varies with @xmath18 and @xmath12 ) and then falls off as a power law according to eq . ( [ powerlaw ] ) with @xmath5 . this power - law behavior appears to be robust in the regime of small @xmath12 and @xmath18 . if , for example , instead of eq . ( [ exponential ] ) we employ a gaussian stress distribution then , although the average distribution of thresholds ( figure [ thresh ] ) changes radically , the power - law form of the avalanche distribution remains . notice , however , that the exponent @xmath1 changes slightly as the applied stresses are varied . for the gaussian distribution , for example , we find @xmath23 , as opposed to @xmath24 for the exponential . and for steeper distributions of stresses ( @xmath25 with @xmath26 ) we find @xmath27 or greater . in order to investigate possible connections with spatially - organized models , we have also implemented our model on a lattice and at each time - step eliminated not only those agents whose thresholds for stress fall below the selected level , but also their neighbors . in all cases we observe a power - law distribution of avalanches with exponent in the vicinity of @xmath28 . in order to understand the appearance of this power law , let us consider the time - averaged behavior of the model . the statistically stationary state arises as a competition between the two processes comprising the dynamics : the stresses which tend to remove lower thresholds from the distribution and thus shift the weight of the distribution to higher values of @xmath29 , and the aging , which tends to move weight back down again . the result is that the average threshold distribution @xmath30 is a highly nonhomogeneous , monotonic increasing function of @xmath29 which , for small @xmath18 , tends to have a plateau as @xmath29 approaches unity ( see figure [ thresh ] ) . by balancing the two competing processes , we can calculate @xmath30 and hence the avalanche distribution . for concreteness , we perform the calculation here for the exponentially distributed stresses of eq . ( [ exponential ] ) . the probability of an agent possessing a threshold @xmath29 lying below the stress level @xmath14 at any given time - step ( and hence of it moving during this time - step ) is @xmath31 the total time - averaged rate at which agents move in the interval between @xmath29 and @xmath32 is then @xmath33 where the @xmath29-independent constant @xmath34 on the right - hand side is the time - averaged rate at which probability is added to @xmath35 . rearranging we have @xmath36 the constant is easily fixed by requiring that @xmath30 integrate to unity , giving @xmath37^{-1}. \label{normalization}\ ] ] for small @xmath12 and @xmath18 , @xmath30 rises exponentially from zero and then levels off in a plateau around @xmath38 . physically , this arises because agents possessing thresholds above this point are affected only by the aging process , which treats them all equally . below this level , the stress process is important too , and it preferentially moves those with lower thresholds . the avalanche size distribution is given by @xmath39 the probability @xmath40 of getting an avalanche of a certain size given a certain stress level , depends on the distribution of thresholds , which will in general vary from one time - step to another . however , if we make the `` time averaged approximation '' ( taa ) whereby one assumes that at each time - step the threshold distribution can be approximated by its time - averaged value , then @xmath41 where @xmath42 is just @xmath43 the avalanche size distribution then becomes @xmath44 where we have used eqs . ( [ pthresh ] ) and ( [ sofeta ] ) . we can calculate the stress level @xmath45 corresponding to an avalanche of size @xmath16 from the same two equations , which give @xmath46 \bigg/ \biggl[\log{1+f{\rm e}^{1/\sigma}\over1+f}\biggr]\nonumber\\ & \approx & \sigma\log ( 1+f{\rm e}^{\eta/\sigma})-\sigma f \label{sapprox}\end{aligned}\ ] ] for @xmath47 and @xmath48 . we can now distinguish a number of different regimes . for small avalanches , such that @xmath49 , the logarithm on the right - hand side can be expanded giving @xmath50 . substituting into eq . ( [ paval ] ) @xmath51^{-2}\qquad\mbox{for $ s\ll\sigma$.}\ ] ] this gives a flat avalanche distribution for small @xmath16 up to about @xmath52 , and then a power - law distribution for larger @xmath16 with exponent @xmath28 . the approximation breaks down when @xmath53 , giving way to a regime in which @xmath54 , and hence the avalanche distribution falls off exponentially with @xmath16 . the various regimes can clearly be seen in the numerical results presented in figure [ size ] , and the predicted cross - over points between them agree well with the theory . when @xmath12 decreases below @xmath55 , the approximations in eq . ( [ sapprox ] ) break down and instead it becomes valid to write @xmath56 . in this regime the theory predicts a breakdown in the scaling , a phenomenon which is also seen in the simulations . thus the reloading process , whose scale is set by @xmath12 , must be small but necessarily non - zero if we are to see power - law behavior in the avalanche distribution . notice however that at precisely @xmath57 the theory predicts a return to @xmath28 scaling , which is not seen in the simulations , implying that the taa breaks down in this regime because the distribution @xmath40 becomes too broad to be well approximated by a @xmath58-function . the physical principle behind the appearance of a power - law distribution here is the interdependence of the avalanche and threshold distributions ; the avalanche distribution is a function of the particular distribution of thresholds at any time , but the threshold distribution is itself produced by the action of the avalanches . there are clear similarities between our model and the sand pile model , in which sites also possess a certain threshold stress that they will withstand without adjusting . furthermore we have a source term , the reloading or aging fraction @xmath12 of agents which at each time - step loose memory of their previously assigned thresholds . this source term is similar in effect to the addition of the single grains of sand in the sand pile models . there are however some important differences between our model and the soc models . first , the stresses in our model are coherent , rather than localized as they are in the sandpile . second , the agents are , at least in the simplest versions of the model , entirely non - interacting . in soc models , it is the interactions which give rise to avalanches . in our model on the other hand the avalanches of simultaneously moving agents arise because all the agents feel the same externally imposed stresses . there is no causal connection between the events which comprise an avalanche ; each agent moves independently of the others . unlike other model systems for large scale fluctuations , such as the burridge - knopoff ( bk ) model @xcite and the recycled version of the democratic fiber bundle model ( dfbm ) @xcite , the model presented here does not make a clear distinction between small , finite - sized events , and large ones whose size scales like the size of the system . in the bk model , for instance , the spectrum of event sizes contains two separate parts , one composed of small events which scales as @xmath59 , and another composed of the big events , which occur quasiperiodically . the bk and dfbm models are not statistically stationary , by contrast with our model whose dynamics rapidly reaches a statistically stationary state . models such as bk and dfbm also show `` foreshock '' events in which large avalanches are preceded by smaller ones . our dynamics does not have foreshocks but does display aftershock events , a phenomenon which we discuss in greater detail in next section . next , we would like to examine the potential relationship of our model to processes occurring in real physical systems . first we consider earthquakes . to begin with , we ignore spatial correlations and consider the variables @xmath10 to be thresholds for movement at various points along a fault . the coherent stress @xmath14 is provided by long - wavelength background noise from some external source , such as other distant tremors , or movements in the deeper regions of the earth , and the reloading @xmath12 is due to slow plastic deformation from tectonic movements of the crust . as we have seen , these elements alone lead directly to a power - law distribution of earthquake sizes very close to the observed gutenberg - richter law , without the need to invoke interactions between neighboring parts of the fault . that is not to say that such interactions do not exist , only that they are not necessary to produce the observed power law . ( kagan @xcite has presented evidence of a fractal pattern in the spatial distribution of earthquake activity , which is an indication that interactions are a feature of the dynamics . this however need not lead us to conclude that these local interactions are necessary for producing the observed size distribution of events . ) another interesting feature of our model is that it shows clear aftershocks . the mechanism for these is straightforward . when a large avalanche takes place , a significant fraction of the thresholds in the system are replaced with new , uniformly distributed ones . because of the monotonic increasing form of the threshold distribution , this has the effect of shifting the weight of the distribution downwards , increasing the fraction of agents with low threshold for movement . the result is that subsequent stresses on the system have a larger - than - normal effect , and we see an amplification of the usual level of `` background '' avalanches in the aftermath of a particularly large event . in figure [ after ] , we show a section of a time series of avalanches from one of our simulations , which clearly displays this aftershock effect . notice that if we apply the argument iteratively , we would also expect to see sequences of `` after - aftershocks '' following each of the aftershocks , a behavior which is indeed evident in figure [ after ] . we have also measured the average probability of getting an event of significant size in the aftermath of another large one , and found that for small times its distribution goes approximately as @xmath60 ( figure [ omori ] ) . a similar result is seen in the data from real earthquakes , and is commonly referred to as omori s law @xcite . the @xmath60 distribution can be understood as follows . a large avalanche will redistribute the thresholds of a large fraction of the agents in the system uniformly across the interval of allowed values ( @xmath61 in this case ) . a subsequent stress of magnitude @xmath62 will remove all those agents with @xmath63 , and produce an aftershock extinction of a certain magnitude . in order to get another significant aftershock we now need a stress @xmath64 in order to reach those agents which were not affected by the first aftershock . in general , if it took a time @xmath65 to get the first stress , then it will on average , take as long again to get another of the same magnitude , or an aggregate time of @xmath66 until a stress of size @xmath67 comes along . repeating the argument , it will take as long again , or a total time of @xmath68 to get the third aftershock , and so forth . given this exponential increase in the time intervals between these events , it is not hard to show that the histogram of aftershock events should have a @xmath60 power - law form , regardless of the precise distribution of stresses applied to the system . note that our mechanism is by no means the only way to obtain aftershocks . an alternative mechanism has been proposed by nakanishi @xcite using a burridge - knopoff - like model in which relaxation processes are introduced by considering the geometry of stress redistribution following large quakes . as with the bk model , nakanishi s model has a quasiperiodic dynamics . second , let us compare our model with the results of recent studies of one - dimensional rice piles by frette _ et al . _ @xcite . in these studies the experimenters found a frequency distribution of avalanche sizes @xmath16 which was flat up to a certain fraction of the total size of the pile , and then fell off as a power of @xmath16 for larger avalanches according to eq . ( [ powerlaw ] ) , with a measured exponent of @xmath6 . a similar behavior is seen in the simulation results from our model ( figure [ size ] ) which also display a flat distribution of avalanches up to a certain fraction @xmath69 of the total system size , and then a power - law fall in avalanche frequency for larger sizes with exponent close to two . a possible interpretation of the experimental data then is that the dynamics of the rice pile is one of avalanches produced by the interplay of reloading with coherent stress . the reloading @xmath12 could arise as a result of newly added grains of rice , which tend to randomize the thresholds for grains on the surface , and the stresses might come from the tumbling of new grains as they are added to the pile . the plateau in the avalanche distribution for small sizes @xmath16 is then caused by rice grains which tumble past a number of sites before coming to rest , but have only enough energy to disturb the most unstable of those sites , and the larger events which form the @xmath28 power law are the result of occasional larger stresses in the tail of the distribution . the one - dimensional nature of the system ensures that all input disturbances propagate through a large portion of the system , and thus may be treated as coherent . in a two - dimensional system this would not be the case , and the pile might well show entirely different dynamics , either possessing a shallower power - law distribution @xmath70 , indicating perhaps that a true soc dynamics is at work , or not possessing a power - law distribution at all , indicating that coherent driving forces are the only mechanism responsible for power laws in this system . our model also makes quantitative predictions about the scaling of the line between the two regimes in the avalanche distribution : the position of the line should go like @xmath71 , the factor of the system size @xmath8 appearing when we shift from measuring avalanches as fractions of the system size to measuring the total energy they release . scaling of precisely this form with @xmath8 is indeed seen in the experiments . frette _ et al . _ also mention that simple scaling disappears when the experiment is repeated with `` rounder '' rice . we can explain this result in terms of the narrower distribution of thresholds that round rice can support , which corresponds to larger values of both @xmath18 and @xmath12 . the results of these experiments have also been modeled by christensen @xcite using a soc model with interacting elements . clearly , there are aspects of the dynamics captured by their model which are missing from ours , particularly geometrical effects concerned with the spatial distributions of avalanches and the corresponding transport properties of rice in the pile @xcite . however , because the exponent @xmath1 is greater than 2 , making @xmath72 independent of system size , we can expect these properties to be independent of the largest avalanche events ( though on the other hand , they should now depend strongly on the position and nature of the crossover between the two regimes of the avalanche distribution ) . we suggest that the reverse is also true , i.e. , that the observed large avalanches could appear even in the absence of long range spatial correlations . one characteristic which does seem to distinguish our model from the soc alternatives is the existence of aftershock events . it might therefore might be profitable to investigate the existence of aftershock avalanches in the experimental data , in order to make a quantitative distinction between the two classes of dynamics . finally we would like to point out that models of the type introduced here do not constrain @xmath1 to values close to @xmath7 . although the values found with the simple version of the model outlined in section ii all lie approximately in the range @xmath73 , we have investigated other variants on the model which produce values outside this range . one particularly interesting version is one in which we allow for the possibility of there being many different kinds of stress on an agent . we suppose that agent @xmath9 is subject to @xmath74 independent types of stress , and that it has a separate threshold for yielding to each one , making @xmath75 an @xmath74-dimensional vector quantity . one then assumes that all @xmath74 components of @xmath76 are to be replaced with new values every time any one of the types of stress exceeds the corresponding threshold value . in the limit @xmath77 , this model is clearly just the same as the version discussed above , and for higher values of @xmath74 we continue to see a power - law distribution of avalanche sizes , regardless of the nature of the applied stresses . however , the exponent of the power law becomes steeper as the value of @xmath74 increases , and appears to approach 3 as @xmath74 becomes large . ( we have investigated the model numerically up to @xmath78 . ) it is interesting to note that stock market fluctuations show power - law fluctuation distributions with exponents close to @xmath79 @xcite . one may speculate whether these so - called `` fat tails '' in the distribution are the natural response to the action of external stresses on the market ( of which there are indeed many ) . to summarize , we have demonstrated that coherent noise in large systems typically gives rise to intermittent behavior with an `` avalanche '' type dynamics characterized by a power - law distribution of avalanche sizes with exponent in the vicinity of @xmath28 . this value is similar to that seen in a number of real systems , including rice piles and earthquakes , suggesting that these systems may in fact be driven by external noise , rather than self - organizing under the influence of short - range internal interactions . if one allows more elaborate types of stress on the system one can obtain power laws with exponents as high as @xmath79 . we believe that the study of systems driven in this fashion by coherent external noise may offer new interpretations of intermittent dynamics in a variety of extended non - equilibrium systems in terms of a direct interplay between small scale structures and long wavelength fluctuations in the system . such systems might include not only the ricepiles and earthquakes considered here , but possibly also extended chaotic systems such as economics @xcite and turbulence @xcite . k. christensen and z. olami , j. geophys . 97 * , 8729 ( 1992 ) ; phys . a * 46 * , 1829 ( 1992 ) . in fact it could be argued that @xmath28 is the upper `` critical '' value for 1d soc models because the spatial organization on a length scale @xmath80 is due to avalanches of size larger than @xmath80 . the probability for these to appear is of order @xmath81 , and , since they affect @xmath80 sites their weighted contribution to ordering on length scale @xmath80 is @xmath82 . this average organization on length scale @xmath80 should be compared with the disorganization caused by non - overlapping avalanches appearing with frequency @xmath83 . criticality demands that organization should be larger than disorganization for @xmath84 , implying that @xmath70 . .	we present a simple model of a dynamical system driven by externally - imposed coherent noise . although the system never becomes critical in the sense of possessing spatial correlations of arbitrarily long range , it does organize into a stationary state characterized by avalanches with a power - law size distribution . we explain the behavior of the model within a time - averaged approximation , and discuss its potential connection to the dynamics of earthquakes , the gutenberg - richter law , and to recent experiments on avalanches in rice piles .
interest in the search for additional information in the hadron component of eas at mountain altitude in terms of very high energy interactions @xcite has been spurred by the necessity to have a continuity of multiproduction features between present colliders at @xmath3 1.8 tev and the future lhc high energy physics at @xmath4 = 16 tev ( @xmath4 represents here the centre of mass energy for the nucleon nucleon collision ) . we review here the data obtained from calorimeters , or equivalent detectors , on high energy hadrons ( longitudinal development and lateral spread ) coupled with air shower arrays , where the primary energy is estimated from the electron and the muon components . from qualitative considerations , the most energetic hadrons ( @xmath5 ) can be supposed to conserve some significant properties correlated with the dynamics of the earliest interactions : for instance , the relation between the transverse momentum @xmath6 , the height h of the interaction , the hadron energy @xmath7 and the radial distance @xmath8 , @xmath9 has been suggested as a possible measure of @xmath6 @xcite . + following this line of inquiry , special attention has been given to the measurements of quantities like @xmath10 , @xmath11 , @xmath0 ( where @xmath12 are the mean values per hadron see @xcite ) and @xmath13 , the number of hadrons above the energy threshold @xmath14 . a convergent and significant increase of @xmath1 was observed in the different experiments @xcite and the earliest interpretation was a tremendous rise of the transverse momentum @xmath15 , above @xmath4 = 5000 gev , the minimum for @xmath6 ( occuring from equation 1 when the hadrons come directly from the first interaction ) being situated around 2 gev / c @xcite . + in more recent analysis measurements of hadron lateral distribution in tien shan calorimeter ( which unfortunately have never been repeated @xcite ) have been interpreted in terms of increasing role of high @xmath6 processes @xcite . the problem still remains open , since the number of reports on the subjest is small and measurements of hadrons suffer more than other eas components from systematic errors which have to be carefully estimated , using simulations of the detector response " @xcite . in parallel with our review of the experimental data , we have carried out eas simulations with the program corsika version 4.50 @xcite , regarded as containing the major features of the present collider physics , in the energy range 10@xmath16 10@xmath17 gev . our analysis has concentrated on the hadron lateral spread to explain the specific effects on the different factors combined in @xmath10 , such as the sum of energy deposited , @xmath18 , the hadron distance , @xmath8 , and the hadron content @xmath13 ( with consequences for the containment as well as the resolution of the detector ) . in order to appreciate the complexity of the hadron cascade mechanism , and how it can conserve or smear out the original interaction characteristics , the progression of our analysis is organized as follows : * physical dependence of interaction parameters ( multiplicity , inelasticity , @xmath6 distribution .... ) * dependence on primary spectrum and composition * dependence on data acquisition and experimental resolution . the survey of this dependence has focused on the largest detector employed , i.e. the tien shan calorimeter of 36 m@xmath19 @xcite . the angular emission of secondary hadrons depends on the ratio @xmath20 , where @xmath21 is longitudinal momentum in laboratory system . its evolution from cm frame to laboratory system is dominated , after lorentz transformation , by the well known collimation effect " , the emitted jet becoming narrower as the hadron energy increases . the multiplicity grows as @xmath22 , inelasticity changes slowly and @xmath6 rises regularly @xcite , but in reasonable proportion ( circle shape points in figure [ fig : ptve ] ) . + other effects , such as the increase in @xmath6 for p a collisions or a a collisions ( as described for the cronin effect " by schmidt and schukraft @xcite ) , or correlated with jet production , or the increase in @xmath6 for diffractive events @xcite are not very important or can be disregarded according to their respective cross sections . the correlation between @xmath15 and multiplicity ( or central rapidity density @xmath23 ) can not be neglected in the treatment of individual events more correlated with the semi inclusive data ; the correlation of the ua1 experiment , described in @xcite , is included in corsika package ( for the option hdpm ) and has been used , without change , in the main part of our simulation . we have , nevertheless , performed some simulations with a @xmath6 generator following the most recent correlation @xmath15 with @xmath23 from the cern ua1 mimi collaboration @xcite , characterized by a greater rise of @xmath15 at large multiplicity . in this way , we ascertained that the subsequent relative changes in @xmath24 are of second order when compared to the attempt to explain @xmath24 with @xmath15 values exceeding 2 gev / c at @xmath4 = 5 tev @xcite ( squares on figure [ fig : ptve ] ) . + furthermore some simulations have been done with this extremal value + @xmath15 = 2 gev / c around a primary energy @xmath25 = 10@xmath26 gev , after a special tuning of the @xmath6 generator in corsika : the options currently used have been hdpm in the usual form @xcite , as well as coupled to the @xmath6 generator elaborated following ua1 mimi results @xcite , and also the venus option @xcite . a detailed description of the evolution with energy of the global interaction parameters used in the options of corsika can be found in references @xcite @xcite . + the physical area explored following our selection of input parameters concerns primary energies between 10@xmath16 10@xmath17 gev , zenith angles between 0 and 30 deg for primary masses a = 1 , 4 , 14 , 28 , 56 . the altitude for observation levels has been set to 3200 m with statistical samples of 500 simulated cascades per primary . showers induced by heavy primaries have been simulated simultaneously with venus option and hdpm combined with superposition or , alternatively , with the abrasion - evaporation procedure @xcite , describing the fragmentation of heavy projectiles . the results obtained for @xmath27 are shown in the figure [ fig : ervne ] ( dashed line ) , where @xmath24 is plotted versus electron size @xmath28 , for proton primaries and for a hadron energy threshold @xmath14 = 1 tev . the expected decrease versus size , as in previous calculations @xcite , is obtained . the agreement is tolerable only for the experimental data @xcite at @xmath28 near @xmath29 . the lateral hadron densities @xmath30 have been described by the experimenters with a simple exponential law : @xmath31 we have used the same functional dependence to fit the hadron distribution in each individual shower with two free parameters : normalization related to the total number of hadrons @xmath32 and mean radius @xmath33 . the total number of hadrons @xmath34 varied from 0.05 for hdpm model and primary fe to 0.9 for venus option and primary proton for @xmath28 = 10@xmath16 and respectively from 3.5 to 9.0 for @xmath28 = 10@xmath35 . the average mean radius as a function of the shower size is presented in figure [ fig : rvne ] . similar remarks as in figure [ fig : ervne ] are suggested by figure [ fig : rvne ] , where @xmath36 decreases versus size , opposite to the general experimental tendency . + @lllll factors & a & @xmath24 & @xmath37 & comment + & & ( tev m ) & ( m ) & + a ) primary mass a & 1 & 3.16 & 1.6 & hdpm + + at @xmath38gev & 4 & 3.4 & 1.70 & abrasion- + & 14 & 3.67 & 2.21 & evaporation @xcite + & 28 & 4.29 & 2.52 & + ua1 mimi + & 56 & 4.7 & 3.24 & 96 @xcite + & & & & + & 56 & 3.81 & 1.34 & hdpm+superp . + & & & & + & 1 & 3.73 & 1.09 & venus + & & & & + & @xmath14 & @xmath24 & @xmath37 & comment + & ( tev ) & ( tev m ) & ( m ) & + b ) hadron & 0.32 & 1.988 & 1.24 & hdpm + threshold energy @xmath14 & 1.0 & 3.34 & 1.01 & + ( primary protons & 3.16 & 5.51 & 0.765 & + of @xmath39gev ) & 10.0 & 8.29 & 0.57 & + & & & & + & @xmath40 & z & @xmath24 & @xmath37 + & ( @xmath41 ) & & ( tev m ) & ( m ) + c ) initial conditions , & 5 & 2 & 5.3 & 3.08 + multiplicity & 5 & 1/2 & 5.6 & 2.78 + ( @xmath42 ) & 150 & 2 & 2.4 & 0.98 + and depth ( @xmath40 ) of first & 150 & 1/2 & 2.5 & 1.26 + interaction ( primary & & & & + protons of @xmath39gev ) & & option : & hdpm & + & @xmath25 & @xmath24 & @xmath37 & option + & ( gev ) & ( tev m ) & ( m ) & + d ) primary energy & 10@xmath16 & 4.248 & 1.122 & hdpm + ( for protons & 3@xmath4310@xmath16 & 3.678 & 1.08 & + and for @xmath14 = 1 tev ) & 10@xmath35 & 3.58 & 1.03 & + & 3@xmath4310@xmath35 & 3.34 & 1.01 & + & 10@xmath26 & 3.00 & 0.95 & + & 10@xmath17 & 2.69 & 0.93 & + the total number of hadrons @xmath13 reflects also a disagreement , characterized by an excess in the simulation @xcite , increasing with primary energy : the amplitude of this disagreement , when compared to the influence of primary energy and interaction parameters , is illustrated by the values calculated in table 1 . + setting @xmath15 above 2 gev / c around a primary energy @xmath25 = 10@xmath26 gev gave no rise in @xmath24 and @xmath37 : a slight decrease was even observed due to the larger transverse mass per secondary and to the consequent enhancement of the inelasticity in the former collision , reducing in proportion @xmath44 at mountain level . + table 1 represents a small part of the total simulation which contains similar effects to the examples presented . the total sample , up to 10@xmath45 ev , confirms general aspects of the dependence , that we can summarize as follows : + * @xmath37 and @xmath24 decrease with energy even if @xmath15 is rising with energy . * this effect persists even for extremal value of @xmath15 = 2 gev / c and the collimation effect " dominates over all the energy interval considered . * the height of the first interaction has an effect on the lateral spread , increasing @xmath24 . * the factors @xmath37 and @xmath24 increase with the primary mass a. the conditions of reception of the main hadron detectors employed at mountain altitude are compared in table 2 . + .hadron detectors at mountain altitude @xcite . [ cols= " < , < , < , < , < , < , < " , ] the hadron energy in the tien shan ionization calorimeter , composed of 850 g @xmath46 lead , is measured at a depth between 130 850 g @xmath46 with an accuracy of 20@xmath47 @xcite . this thickness represents only 3.7 @xmath48 ( for an interaction length in lead @xmath48 = 194 g @xmath46 ) corresponding to a containment of about 80@xmath47 of the cascades induced by individual hadrons @xcite . the coordinates of the single hadrons are measured with an accuracy of 25 cm in the 744 ionization chambers of the calorimeter : those chambers , with an area of ( 25 @xmath49 300 ) cm@xmath19 , originally arranged in 15 rows , are intended to measure the hadron number with an accuracy of 5@xmath47 @xcite . correction coefficients taking into account the limited lateral resolution and the possibility of aggregation of several hadrons within the chamber size bin ( 0.25 @xmath49 3 ) m were introduced by danilova @xcite to recalculate the hadron energy spectra . the original method to separate and count the hadrons was based on the histograms of the ionization energy distribution in different rows in x and y @xcite . + the limited efficiency of this treatment can be seen in figure [ fig : exam ] , which shows real contained target diagrams for hadrons above 1 tev and the same events reconstructed with the tien shan procedure ; other difficulties appear also in the geometrical containment , the fraction of hadrons registered in one shower decreasing strongly with the primary energy ( from 60@xmath47 at @xmath51 down to 20@xmath47 at @xmath52 according to our simulation ) . + the events simulated for section 2 have been treated by the numerical procedure of reconstruction of experimentally recorded detector response . the energy of each hadron was added to a respective ionization chamber output . as it is shown in figure [ fig : exam ] from a detector plane four one - dimensional projections are obtained ( each chamber is 3 m long while the detector size is 6 m @xmath49 6 m ) . then on each projection the possible hadronic cluster `` coordinates have been determined . using as a separation criterion the dip in the projected energy deposit distribution of at least 70% of neighbouring chamber outputs , close chamber signals could be combined just as it has been performed in the tien shan experiment . then after comparing the amplitudes on respective x- and y - projections taking into account the possibility of overlapping of some hadrons in one of the projections the so - called hadronic clusters '' have been found . they will be called hereafter the reconstructed hadrons . they are shown in the right plots in figure [ fig : exam ] . + + our reconstruction procedure includes also hadrons of energies below the threshold ( 1tev ) which are mostly added to the large hadronic clusters . the correction coefficient of 2.64 ( equal exactly to the one used by the tien shan group ) have been applied to the cluster energy while converted to the reconstructed hadron energy . the position of the reconstructed hadron was the energy weighted mean of chamber outputs in each projection . as the shower axis coordinates we have used the most energetic reconstructed hadron coordinates just like it has been done in tien shan experiment . + for all situations considered in section 2 the general decrease for @xmath24 and @xmath53 vs size turns to a large increase ( solid line in figure [ fig : ervne ] and figure [ fig : rvne ] ) , as well as @xmath24 vs @xmath14 ( figure [ fig : ervet ] ) . this is confirmed ( figure [ fig : latdis ] ) by comparison with the most recent data @xcite where the individual reconstruction indicates how the hadron content is underestimated near the axis . to preserve the clarity of the graphs , we have not plotted on figures [ fig : ervne ] and [ fig : rvne ] the results of the experiments @xcite , @xcite , @xcite @xcite confirming the tendencies shown . those values normalized to tien shan data can be found in reference @xcite and reference @xcite . the proportion between the factor @xmath37 calculated in table 1 and the resolution distance quoted in table 2 shows the role of underestimation of hadron densities near the axis ( when several hadrons enter simultaneously in the same chamber bin ) . this effect will be even greater than in the tien shan ionization calorimeter in all the other detectors quoted . the values of @xmath14 correspond to larger hadron densities near the axis , increasing the probability of aggregation in the same chamber bin . also for larger resolution distance ( i.e. wider bins ) this probability increases . + the simulation suggests that the hadron component and its lateral spread needs to be measured with more accurate calorimeters and a resolution distance substantially smaller than 25 cm . + we have shown that the rise @xmath24 is a consequence only of the complex treatment of the data which is inadequate for the superposition of hadrons , and the rise of @xmath15 derived from @xmath24 and shown in the figure [ fig : ptve ] is not valid . the lateral spread of muons above 200 and 500 gev , also correlated with @xmath15 in earliest interactions , does nt require such an enhancement of @xmath15 @xcite ; the main bulk of these lateral muon distributions being explained with a constant value of @xmath15 = 0.4 gev / c @xcite . as follows from our simulations , even if @xmath15 is rising with energy , the hints in @xmath37 and @xmath24 can not be expected to reflect the @xmath6 behaviour , as they depend on several factors of the hadron cascade developement . to explain the enhancement in figure [ fig : ervne ] , an extremely large increase in primary mass , proportional to the primary energy , was introduced in the simulation @xcite to transform the general decrease of @xmath24 versus @xmath28 in a visible increase . such conjectures are no longer necessary and the agreement can be obtained without a large increase of the primary mass . + the @xmath6 dependence used here in the 3 options of corsika covers a wide range of models ; for instance , the values of @xmath6 in option 2 are very close to the values used in sybill @xcite as shown in @xcite by knapp , heck and schatz , and we can conclude that the present measurements of hadrons at mountain altitudes are in agreement with a majority of models used in eas simulations . they require neither a large @xmath15 enhancement nor a noticeable increase of primary mass . their significance will remain limited until the resolution distance of the detectors is seriously improved . as it was advanced by clay @xcite , the results in hadron measurements depend critically on how well individual hadron signals are resolved in the hadron detector " and this has been confirmed by the present work . nikolsky s i 1998 _ proc . 15@xmath54 ecrs ( perpignan ) _ , nuclear physics b ( proc . suppl . ) vol 60b p 144 winn m m 1977 _ proc . 15@xmath54 icrc ( plovdiv ) _ vol 10 p 305 romakhin v a , nesterova n m , dubovy a g 1977 _ proc . 15@xmath54 icrc ( plovdiv ) _ vol 8 p 107 nesterova n m and romakhin v a 1977 _ proc . 15@xmath54 icrc ( plovdiv ) _ vol 8 p 113 aseikin v s et al . 1981 _ prepr . no 178 _ ( lebedev inst . , nauk , moscow ) p 4 . sreekantan s v , tonwar s c and viswanath p s 1983 _ phys . d _ vol 28 p 1050 bhm e and g.holtrup g 1974 _ core structure of the hadron component in eas , 4@xmath54 ecrs ( lodz ) _ preprint nesterova n m and romakhin v a 1998 _ private communication _ dubovy a g , nesterova n m , chubenko a p 1991 _ journal of nuclear physics _ vol 54 p 178 sivaprasad k 1996 _ il nuovo cimento _ vol 19c no 5 p 643 ( rapporteur talk given at the 24@xmath54 icrc , rome , august 28 sept . 8 , 1995 ) knapp j , heck d and schatz g 1997 _ fk report 5828 _ capdevielle j n et al . 1992 _ the corsika simulation program , kfk report , ( karlsruhe ) _ schmidt h r and schukraft j 1993 _ _ vol 19 p 1705 goulianos k 1983 _ phys . _ vol 101 p 169 bocquet g et al . 1996 _ phys.lett . b _ vol 366 p 434 ( ua1 - mimi collaboration ) attallah r , capdevielle j n , meynadier c , szabelska b , szabelski j 1996 _ _ vol 22 p 1497 machavariani s k , nesterova n m , nikolsky s i , romakhin v a , tukish e i 1981 _ proc . 17@xmath54 icrc ( paris ) _ vol 6 p 193 romakhin v a and nesterova n m 1979 _ acad . nauk , trudy ord . lebedev ( moscow ) _ vol 109 p 77 capdevielle j n et al . 1997 _ proc . 9@xmath54 is vhecri ( karlsruhe ) , nuclear phys . b _ vol 52b p 146 dubovy a g and nesterova n m 1983 _ proc . 18@xmath54 icrc ( bangalore ) _ vol 6 p 82 miyake s , ito n , kawakami s , hayashi y , awaji n 1979 _ proc . 16@xmath54 icrc ( kyoto ) _ vol 13 p 165 matano t , ohta k , machida m , kawasumi n , tsushima i , honda k , hashimoto k , aguirre c , anda r , navia c 1979 _ proc . 16@xmath54 icrc ( kyoto ) _ vol 13 p 185 ticona r et al . 1993 _ proc . 23@xmath55 icrc ( calgary ) _ vol 4 p 331 nesterova n m and dubovy a g 1979 _ proc . 16@xmath54 icrc ( kyoto ) _ vol 8 p 345 bintinger d 1989 _ proc . workshop on calorimetry for supercollider ( tuscaaloosa ) _ r.donakson , world scientific teaneck n.j . ) p 91 danilova t v , dubovy a g , erlykin a d , krutikova n p , nesterova n m , nikolsky s i , yakovleva t i 1987 _ proc . 20@xmath54 icrc ( moscow ) _ vol 6 p 47 vashkevitch v v et al . 1988 _ jad . _ vol 47 p 1054 khrenov b a 1988 _ proc . 5@xmath54 is vhecri , inv . and rapporteur papers _ ( university of lodz publishers ) p 151 gaisser t k et al . 1989 _ phys . _ vol 62 p 1425 clay r w 1985 _ proc . 19@xmath54 icrc ( la jolla ) _ vol 9 p 357 ( rapporteur paper )	an extensive simulation has been carried out to estimate the physical interpretation of dynamical factors such as @xmath0 , @xmath1 in terms of high energy interaction features , concentrated in the present analysis on the average transverse momentum . + it appears that the large enhancement observed for @xmath1 versus primary energy , suggesting in earliest analysis a significant rise of @xmath2 with energy , is only the result of the limited resolution of the detectors and remains in agreement with a wide range of models used in simulations . [ high energy hadrons in eas ]
neutrino - driven winds from core - collapse supernova explosions contribute to the synthesis of elements beyond iron . after the explosion , the hot proto - neutron star cools emitting neutrinos . these neutrinos interact with the stellar matter and deposit energy in the outer layers of the proto - neutron star leading to a supersonic outflow known as neutrino - driven wind @xcite . although neutrino - driven winds were considered the site where heavy elements are produced by the r - process @xcite , recent simulations @xcite can not reproduce the extreme conditions required for producing heavy r - process elements ( see e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? the wind entropy is too low ( less than @xmath12 ) and , even more significant , the ejecta is proton rich ( the electron fraction @xmath13 remains above 0.5 during seconds , see @xcite ) . even if the r - process does not take place in every neutrino - driven wind , lighter heavy elements ( e.g. , sr , y , zr ) can be synthesized in this environment as suggested by @xcite . in proton - rich conditions @xcite showed that elements beyond @xmath14ge can be synthesized . @xcite found sr , y , zr in small pockets of neutron rich material ejected after the explosion of low mass progenitors . recently , @xcite performed a systematic nucleosynthesis study that strongly supports the production of lighter heavy elements in proton- and neutron - rich neutrino - driven winds . in proton - rich winds , charged particle reactions ( alpha and proton captures ) build nuclei up to @xmath10ni and even up to @xmath14ge once the temperature drops below 3 gk . due to their long beta - decay lifetimes and low thresholds for proton capture , the nuclei @xmath10ni and @xmath14ge act as bottlenecks that inhibit the production of heavier elements . in the @xmath0-process , their decay is sped up by @xmath4 reactions , with the neutrons produced by antineutrino absorption on the abundant free protons . this allows for the production of elements beyond iron and may explain the origin of light p - nuclei @xcite . the synthesis of elements by the @xmath0-process depends thus on neutrino spectra and luminosities but also on the dynamical evolution as matter expands through the slow , early supernova ejecta . this produces a wind termination shock or reverse shock where kinetic energy is transformed into internal energy @xcite . the reverse shock has a big impact on the nucleosynthesis because temperature and density increase and the expansion is strongly decelerated . this hydrodynamical feature has been extensively studied for r - process nucleosynthesis @xcite . recently , @xcite have also explored the relevance of the reverse shock on the @xmath15-process . motivated by their work and by new 2d hydrodynamical simulations of the neutrino - driven wind @xcite , we investigate here the wind termination shock to gain further insights on the dynamical evolution relevant for the @xmath0-process . here we use a trajectory from hydrodynamical simulations ( sect . [ sec : dynamics ] ) combined with a complete nucleosynthesis network ( sect . [ sec : netw ] ) . in sect . [ sec : results ] we present our results where we analyze the impact of the wind termination temperature ( sect . [ sec : constt ] ) , the temperature jump at the reverse shock ( sect . [ sec : rs ] ) , and the influence of late temperature evolution and the decay to stability ( sect [ sec : long ] ) . our conclusions are summarized in sect . [ sec : conclusions ] . our study is based on one trajectory obtained from hydrodynamical simulations by @xcite . these simulations follow the supernova explosion and the subsequent neutrino - driven wind . gk and produces a jump in density and temperature and a sudden decrease of the expansion velocity.,title="fig : " ] + figure [ fig : trajectory1d ] presents the evolution of density and temperature for the chosen trajectory . in the following , we always use the same initial evolution and modify it at temperatures below 3 gk . we assume the wind terminates at a temperature @xmath16 and the evolution thereafter is varied using the prescription introduced by @xcite and motivated by 1d and 2d hydrodynamical simulations @xcite . the thermodynamical conditions at the termination of the wind depend on both the intrinsic properties of the wind , like its velocity and mass outflow rate , and on the pressure of the slow moving ejecta . the wind properties are determined by the neutrino energies and luminosities and by the neutron star mass and radius ( see e.g. , * ? ? ? these conditions depend mainly on the nuclear equation of state and are similar for different progenitors @xcite . the properties of the slow supernova ejecta are related to the progenitor and become very anisotropic during the evolution @xcite . in our calculations , we use a simple model that reproduces the main features seen in hydrodynamical simulations . once the wind reaches the early supernova ejecta we use the rankine - hugoniot conditions to determine the behavior of temperature , density , and velocity . when the wind moves supersonically these quantities jump at the reverse shock . the evolution after such discontinuity is determined assuming constant density and temperature during a time @xmath17 . as the mass outflow is constant ( @xmath18 ) the velocity drops as @xmath19 . at later times , the velocity stays constant and the density decreases as @xmath20 ( see @xcite for more details ) . to account for the differences found in the late evolution in two - dimensional simulations , we will use different values of @xmath21 and explore the impact on the nucleosynthesis . the evolution of the composition is calculated using a full reaction network @xcite . all calculations start when the temperature drops below 10 gk and are followed until the temperature reaches 0.01 gk . we assume the initial composition to be determined from nuclear statistical equilibrium for a fixed electron fraction @xmath22 . the reaction network includes 1869 nuclei from free nucleons to dysprosium ( @xmath23 ) . neutral and charged particle reactions are taken from the reaclib compilation , containing theoretical statistical - model rates by @xcite ( non - smoker ) and experimental rates by @xcite . the theoretical weak interactions rates are taken from @xcite and @xcite . where available , experimental beta - decay rates are used from nudat2 supplemented by theoretical beta - decay rates @xcite . neutrino absorption on nucleons and nuclei is also taken into account . we assume a constant neutrino luminosity for neutrinos and anti - neutrinos ( @xmath24 erg ) , and fermi - dirac spectra with a temperature consistent with the hydrodynamical simulations . in order to understand the nucleosynthesis evolution , it is useful to look at the time variations of the mean lifetimes for @xmath1-decays , @xmath11 , @xmath3 , @xmath4 , @xmath25 , and @xmath26 reactions . these mean lifetimes ( or average timescales ) are defined as @xmath27 ^{-1},\ ] ] @xmath28^{-1},\ ] ] @xmath29^{-1},\ ] ] @xmath30^{-1},\ ] ] @xmath31^{-1},\ ] ] and @xmath32^{-1},\ ] ] where @xmath33 denotes the decay rate of nucleus @xmath34 , @xmath35 the reaction rate for process @xmath36 on nucleus @xmath34 , and @xmath37 , and @xmath38 the photo - dissociation reaction rates with emission of neutron and proton , respectively . in these equations , @xmath39 is avogadro number . the average is taken over the heavy nuclei ( @xmath40 ) : @xmath41 another useful quantity to understand the nucleosynthesis evolution is the reaction flow between two nuclei @xmath42 and @xmath43 defined as @xmath44 the quantity @xmath45 denotes the change in abundance of nucleus @xmath42 due to all reactions connecting nucleus @xmath42 with nucleus @xmath43 . the largest reaction flows at each time step indicate the key reactions and thus the nucleosynthesis path . we study here the impact of the wind termination varying the temperature at which it occurs . we assume that the velocity is subsonic and hence there is no jump in temperature , density , and velocity . the different temperature evolutions and their nucleosynthesis are shown in fig . [ fig : tempconst ] . the original trajectory ( solid black line ) is also included for completeness and comparison purposes . in the modified trajectories the wind termination is at @xmath46 1 , 1.5 , 2 , 2.5 , and 3 gk . the resulting abundances change significantly with temperature @xmath16 . there is an optimal @xmath16 around @xmath47 gk for producing heavier elements @xcite . moreover , the nucleosynthesis evolution is very different for temperatures higher and lower than this optimal temperature . this can be seen in fig . [ fig : timescale ] where the relevant averaged timescales ( see sect . [ sec : netw ] ) are shown for the trajectories with @xmath46 1 , 2 , 3 gk . initially , there is @xmath48 equilibrium in all cases . the timescales for the @xmath3 and @xmath26 reactions are much shorter when the wind termination occurs at higher temperatures ( see bottom panel of figure [ fig : timescale ] ) . [ cols="^ " , ] for the wind termination at low temperatures , @xmath46 1 , 1.5 gk , only elements up to germanium are produced in substantial amounts . in these evolutions matter expands very fast . therefore , it only stays for a short time in the temperature rage of @xmath49 gk where charged - particle reactions can effectively synthesize heavier nuclei . in addition , matter rapidly reaches large radii where the antineutrino flux is rather low . similar conditions are found in explosions of low mass progenitors where the expansion of the supernova ejecta is very fast @xcite . notice that the wind termination at a slightly higher temperature , 1.5 gk instead of 1 gk , allows to reach somewhat higher mass number . for the wind termination at high temperatures , one would expect that photo - dissociation stops the synthesis of heavier elements . in this case , the abundance would continuously shift towards lower mass number as temperature increases . however , in our calculations we see an abrupt change in the abundances when the wind termination temperature exceeds certain value . this points to a key reaction acting as a bottleneck at high temperatures . we have identified such a reaction using the reaction flows introduced in sect . [ sec : netw ] . figure [ fig : flux ] shows these flows at two different temperatures for the evolution with @xmath46 2 gk . the upper panel corresponds to @xmath50 gk , and the bottom one to @xmath51 gk , both represent conditions before the constant temperature phase . notice that the reactions that determine the nucleosynthesis flow are different at high and low temperatures . at high temperatures , the reaction @xmath8cu@xmath9@xmath10ni dominates over @xmath8cu@xmath3@xmath6zn while at low temperatures the contrary is true ( see fig . [ fig : cu59 ] ) . when the @xmath8cu@xmath9 reaction dominates , the nucleosynthesis flow is confined into a closed nicu cycle . a similar behavior has been found in the rp - process for the snsbte cycle @xcite . 2.2 gk.,title="fig : " ] + 2.2 gk.,title="fig : " ] cu , with @xmath52 and @xmath53 . at high temperatures @xmath8cu@xmath9@xmath10ni dominates while for lower temperature @xmath8cu@xmath3@xmath6zn dominates . dotted line indicates where rates are equal . ] in the calculation with @xmath54 gk , when temperatures are still above @xmath55 gk ( see fig . [ fig : cu59 ] ) the heaviest nucleus produced is @xmath10ni due to the nicu cycle . once the temperature drops , the cycle opens and the path reaches @xmath14ge . however , during the high constant temperature phase , the triple alpha reaction maintains a continuous production of seed nuclei from light nuclei with @xmath56 @xcite . this leads to a reduction of the proton abundance and thus of the neutrons produced by antineutrino absorption . notice that the ratio of neutrons produced per seed nucleus is a useful guide of the strength of the @xmath15-process @xcite . both effects ( the nicu cycle and the reduced neutron - to - seed ratio ) result in a complete shutdown of @xmath0-process nucleosynthesis . when the wind termination temperature is very high ( @xmath57 gk ) or very low ( @xmath58 gk ) , elements beyond germanium are hardly synthesized however , the final abundances are different for these two extreme cases . when the wind termination takes place at high temperatures ( @xmath59 gk ) matter accumulates in iron group nuclei due to the nicu cycle and to the continuous production of seed nuclei . while for the wind termination at low temperatures matter moves fast to large distances where the antineutrino flux is not large enough to produce a @xmath0-process . in addition , the low temperatures inhibit the production of seed nuclei . therefore , the final proton abundance is significantly higher in the evolution with the wind termination at temperatures @xmath60 gk . when matter moves supersonically there is a jump at the wind termination in temperature and density , also called reverse shock . here , we analyze the impact of such a jump on the @xmath15-process nucleosynthesis using the trajectories shown in fig . [ fig : jump ] . the trajectory without jump corresponds to the evolution with @xmath61 gk presented in the previous section . the trajectory with jump is chosen such that the temperature increases after the wind termination reaches @xmath62 gk ( same temperature as in the trajectory without jump ) . the constant temperature phase is the same for both evolutions ( @xmath63 s ) . however , the final abundances are very different as shown in fig . [ fig : jump ] . + + in the evolution with jump , the expansion continues very fast between the moment the temperature reaches 2 gk ( marked with an a in the temperature curve , fig . [ fig : jump ] ) and the position of the wind termination shock ( marked with b ) . during this phase , there is not enough time for antineutrino absorption on protons to produce the necessary amount of neutrons to reach heavier nuclei . already in this phase , between a and b , matter starts to beta decay towards stability because temperatures become too low for proton - capture reactions . the temperature rise at the wind termination , from point b to c , increases the effectiveness of proton - capture reactions . this results in the matter flow moving again away from stability and towards heavier nuclei . note that the temperature and matter distribution after the wind termination are very similar for both trajectories ( points a and c ) . for the evolution with a jump at the reverse shock . ] the relevant timescales for the trajectory with jump are presented in fig . [ fig : jump_tscale ] . this can be compared to the middle panel in fig . [ fig : timescale ] that corresponds to the trajectory without jump . after the wind termination , proton - captures are faster for the evolution without jump . more significant are the differences in @xmath11 and @xmath4 timescales : both are much shorter in the evolution without jump ( middle panel in fig . [ fig : timescale ] ) . these differences in the relevant timescales have a big impact on the abundances . the reactions involving neutrons , i.e. , @xmath11 and @xmath4 , depend on the neutron density which is shown in fig . [ fig : jump_nn ] for the two evolutions . in the trajectory with jump ( solid line ) , the neutron density remains lower than in the trajectory without jump at all times , even at the wind termination ( feature at @xmath64 s ) . in the @xmath0-process there is an equilibrium between neutron capture and neutron production by antineutrino absorption on protons . this implies that @xmath65 here @xmath66 is the electron antineutrino absorption rate and @xmath67 is the sum of reaction rates for @xmath11 and @xmath4 reactions for nucleus ( z , a ) . therefore , the neutron density in equilibrium is given by @xmath68 the nucleosynthesis path is very similar for both trajectories considered here ( evolution with and without jump ) . therefore , only small variations are expected in the denominator . consequently , the difference in the neutron densities ( see fig . [ fig : jump_nn ] ) is due to the neutron production by antineutrinos . the deceleration of the expansion at the wind termination occurs at a smaller radii for the trajectory without jump ( see bottom panel of fig . [ fig : jump_nn ] ) . the production of neutrons is hence less efficient in the case with jump because matter reaches larger radii where the neutrino flux is reduced due to its @xmath69 dependency . the difference in the abundance pattern for nuclei above mass number @xmath70 ( see fig . [ fig : jump ] ) is due to this difference in the efficiency of neutron production . . the solid black line corresponds to the evolution with jump , the dashed green line to the evolution without jump.,title="fig : " ] + . the solid black line corresponds to the evolution with jump , the dashed green line to the evolution without jump.,title="fig : " ] + we explore the impact of the dynamical evolution after the wind termination . in this section , we assume that the wind termination occurs at @xmath71 gk and we vary the parameter @xmath21 . the quantity @xmath17 characterizes the timescale for the transition from an expansion with constant temperature and density ( during which @xmath19 ) to a constant velocity expansion with @xmath72 . we choose the values @xmath73 0.0 , 0.5 , and 1.0 s , motivated by the anisotropic evolution of the ejecta in 2d hydrodynamical simulations @xcite . there is a smooth transition between the two extreme cases of @xmath74 s and @xmath75 s , which can be seen in the intermediate case of @xmath76 s. the latter ( @xmath76 s ) was used in previous sections . the different temperature evolutions ( top panel ) and the resulting nucleosynthesis ( middle and bottom panels ) are shown in fig . [ fig : dt - y ] . while all three cases produce similar abundance distributions , the details depend critically on the value of @xmath21 . are 0.0 s ( solid black ) , 0.5 s ( dashed green ) , and 1.0 s ( dotted red ) . abundances for these evolutions are shown versus mass number @xmath77 ( middle panel ) and versus atomic number @xmath78 ( bottom panel).,title="fig : " ] are 0.0 s ( solid black ) , 0.5 s ( dashed green ) , and 1.0 s ( dotted red ) . abundances for these evolutions are shown versus mass number @xmath77 ( middle panel ) and versus atomic number @xmath78 ( bottom panel).,title="fig : " ] are 0.0 s ( solid black ) , 0.5 s ( dashed green ) , and 1.0 s ( dotted red ) . abundances for these evolutions are shown versus mass number @xmath77 ( middle panel ) and versus atomic number @xmath78 ( bottom panel).,title="fig : " ] + + + for a relatively long phase of constant temperature ( @xmath79 s ; left column of fig . [ fig : dt - tau - steps ] ) , the @xmath48 equilibrium lasts for quite a long time and allows for many @xmath4 and @xmath3 reactions to occur . this drives the matter to higher mass number ( up to @xmath80 ) , where matter accumulates in sn ( @xmath81 closed shell ) . due to the extended period of @xmath48 equilibrium , significant abundances of nuclei with @xmath82 are synthesized ( see line c in fig . [ fig : dt - tau - steps ] left column ) . also during this phase there is a continues production of seed nuclei at the expenses of alpha particle ( see fig.11 ) . once the temperature drops below @xmath83 1.5 gk ( line d ) , the production of seed stops . in addition , proton - capture reactions become hindered by the coulomb barrier and @xmath1-decays become faster than @xmath4 reactions ( see upper left panel in fig [ fig : dt - y ] ) . the nucleosynthesis ceases to efficiently proceed to higher masses . during this phase , matter starts to decay towards stability . a small number of late time neutron capture reactions smooths the abundance distribution in the mass range @xmath84 ( line e ) . once the temperature drops below 1.0 gk ( lines e and f ) , the abundances as function of mass number do not change anymore , as at this time @xmath1-decays dominate . the closed shell at @xmath81 acts as barrier for the nucleosynthesis . this can be seen in the enhanced abundances at @xmath85 ( originating from @xmath86sn ) in the final abundance distribution . for the case @xmath75 s ( right column in fig . [ fig : dt - tau - steps ] ) , the situation is very different . the initial phase of @xmath48 equilibrium is much shorter than in the case of @xmath74 s. in this case nuclei up to mass number @xmath80 are synthesized ( line c ) the temperature has already dropped below 2 gk . due to the faster temperature decline , the production of seed nuclei is less efficient than in the case of @xmath74 s ( see fig . [ fig : yn - ya - yh ] ) . this leads to lower abundances of nuclei such as @xmath87c , @xmath88o , @xmath89ne , @xmath90si , and @xmath91ca for @xmath92 s ( see fig . [ fig : dt - y ] ) . the lower abundance of seed nuclei leads to a higher neutron density ( see eq . [ eq : n - equil ] ) as in both cases the neutron production rate ( @xmath93 ) is similar . therefore , in the evolution with @xmath92 s , there are more neutrons available per seed nuclei ( fig . [ fig : yn - ya - yh ] ) . the neutron capture reactions , i.e. @xmath11 and @xmath4 , become faster than @xmath3 reactions at temperatures of @xmath94 gk . these neutron captures move matter to the neutron - rich side of stability . this , together with the earlier less efficient production of seeds , results in a depletion of the abundances in the region of @xmath82 ( compare black solid and red dashed line in fig . [ fig : dt - y ] ) . towards the end of the evolution , @xmath11 reactions are the fastest ones , even faster than @xmath1-decays . during this phase , the atomic number @xmath78 remains unchanged while the mass number increases ( lines d , e , and f in fig . [ fig : dt - tau - steps ] ) . the signature of matter moving to stability via @xmath11 reactions can be seen in the increased abundances at @xmath95 , corresponding to @xmath96sn , the most neutron - rich stable sn isotope . particles ( top panel ) , abundances of heavy nuclei with @xmath40 ( middle panel ) , and neutron densities ( bottom panel ) for @xmath75 s ( solid black ) , 0.5 s ( dashed green ) , and 1.0 s ( dotted red).,title="fig : " ] particles ( top panel ) , abundances of heavy nuclei with @xmath40 ( middle panel ) , and neutron densities ( bottom panel ) for @xmath75 s ( solid black ) , 0.5 s ( dashed green ) , and 1.0 s ( dotted red).,title="fig : " ] particles ( top panel ) , abundances of heavy nuclei with @xmath40 ( middle panel ) , and neutron densities ( bottom panel ) for @xmath75 s ( solid black ) , 0.5 s ( dashed green ) , and 1.0 s ( dotted red).,title="fig : " ] we have studied the impact of the supernova dynamical evolution on the @xmath0-process and how the wind termination affects the synthesis of elements beyond iron . for a robust and strong production of nuclei with @xmath82 there is an optimal wind termination temperature at @xmath832 gk as it was found by @xcite . also in agreement with their work , we have shown that if the wind termination occurs at low temperatures ( @xmath97 gk ) , matter stays too short time in the optimal temperature range for @xmath0-process nucleosynthesis , @xmath98 gk . moreover , the electron antineutrino flux quickly becomes too small to produce the necessary neutrons to overcome the @xmath99-decay waiting points . this hinders the @xmath0-process and consequently the efficient synthesis of heavy nuclei . therefore , the wind termination temperature determines the heaviest elements produced . we have identified an end point cycle that is key at high temperatures ( around 3 gk ) . the close nicu cycle inhibits the production of elements heavier than 56ni . the reactions in this closed nicu cycle are the following : @xmath100 at high temperatures the reaction @xmath8cu@xmath9 @xmath10ni prevents the synthesis of elements beyond the iron group . when the temperature drops below @xmath57 gk , @xmath8cu@xmath3 @xmath6zn becomes more effective than @xmath8cu@xmath9 @xmath10ni . this leads to breakout from the nicu cycle and the flow of matter can continue towards heavier nuclei . the temperature dependence of these two reactions is critical because it sets the temperature at which the @xmath0-process can start to synthesize elements beyond nickel . if this cross - over temperature were slightly higher , matter would be closer to the proto - neutron star and thus under higher neutrino flux when the path starts to move towards heavier nuclei . this will significantly increase the efficiency of the @xmath0-process . therefore , our results clearly motivate further investigation of these two key reactions : @xmath8cu@xmath9 @xmath10ni and @xmath8cu@xmath3@xmath6zn . particularly relevant are the branching ratios for the decay by alpha and gamma emission of compound states in @xmath6zn above the proton separation energy . we have also explored for the first time the impact of the dynamical evolution after the wind termination on the synthesis of heavy elements by the @xmath0-process . hydrodynamical simulations show that after the wind termination there is a transition from an expansion with almost constant temperature and density to a phase with almost constant velocity . the duration of the constant temperature phase , @xmath17 , strongly affects the final abundances of heavy nuclei . we have found that for @xmath101 s the final flow of matter to stability occurs by neutron capture reactions , i.e @xmath11 and @xmath4 , and not by @xmath1-decays . when the phase of constant temperature is very short ( @xmath102 s ) , the abundances of nuclei with @xmath103 are lower . this leads to significantly higher neutron densities at later times . depending on the neutron density the matter moves to stability either by @xmath1-decays ( @xmath79 s ) or by neutron captures . these two different ways of reaching stability leave a distinct fingerprint in the final abundances . we have investigated in detail the impact of dynamical evolution on the @xmath0-process nucleosynthesis using individual characteristic trajectories . in order to predict the complete @xmath0-process yields from a supernova simulation , one will need to integrate over all proton - rich ejecta . in summary , the supernova dynamics as well as individual reactions such as the proton capture reactions on @xmath8cu determine how high in proton number the @xmath0-process can proceed . the dynamical evolution just after the wind termination can greatly affect the nucleosynthesis evolution towards stability as it determines the late neutron density . our results provide a link between nucleosynthesis in proton - rich winds and the dynamical evolution of the ejected matter and motivate further theoretical and experimental effort on understanding key reactions . we thank f. montes , t. rauscher , and f .- k . thielemann for valuable discussions and u. frischknecht and c. winteler for their support preparing the flux figures . acknowledge support from the alexander von humboldt foundation and the swiss national science foundation . c.f . acknowledges support from the doe topical collaboration `` neutrinos and nucleosynthesis in hot and dense matter '' under contract de - fg02 - 10er41677 . is partly supported by the deutsche forschungsgemeinschaft through contract sfb 634 , the helmholtz international center for fair within the framework of the loewe program launched by the state of hessen and the helmholtz association through the nuclear astrophysics virtual institute ( vh - vi-417 ) .	we study the impact of the late time dynamical evolution of ejecta from core - collapse supernovae on @xmath0-process nucleosynthesis . our results are based on hydrodynamical simulations of neutrino wind ejecta . motivated by recent two - dimensional wind simulations , we vary the dynamical evolution during the @xmath0-process and show that final abundances strongly depend on the temperature evolution . when the expansion is very fast , there is not enough time for antineutrino absorption on protons to produce enough neutrons to overcome the @xmath1-decay waiting points and no heavy elements beyond @xmath2 are produced . the wind termination shock or reverse shock dramatically reduces the expansion speed of the ejecta . this extends the period during which matter remains at relatively high temperatures and is exposed to high neutrino fluxes , thus allowing for further @xmath3 and @xmath4 reactions to occur and to synthesize elements beyond iron . we find that the @xmath0-process starts to efficiently produce heavy elements only when the temperature drops below @xmath5 gk . at higher temperatures , due to the low alpha separation energy of @xmath6zn ( @xmath7 mev ) the reaction @xmath8cu@xmath9@xmath10ni is faster than the reaction @xmath8cu@xmath3@xmath6zn . this results in the closed nicu cycle that we identify and discuss here for the first time . we also investigate the late phase of the @xmath0-process when the temperatures become too low to maintain proton captures . depending on the late neutron density , the evolution to stability is dominated by @xmath1 decays or by @xmath11 reactions . in the latter case , the matter flow can even reach the neutron - rich side of stability and the isotopic composition of a given element is then dominated by neutron - rich isotopes .
although far less common than low - mass stars , massive stars play a key role in the evolution of the energetics and chemistry of molecular clouds and galaxies . however , the formation of high - mass stars is far less clear than their low - mass counterparts for several reasons . massive stars are rare and evolve quickly . in addition , the regions that host their early stages of formation are difficult to observe due to high dust extinction , and , with a few exceptions , are located at large distances ( @xmath4 3 kpc ) . for this reason , most of the research on massive star formation has been based on observations of regions with current star formation ( e.g. , `` hot cores , '' regions ) , which were initially easier to detect in surveys and to follow up in detail . in constrast , objects in the earlier `` prestellar '' or `` starless '' stage have been much harder to find and , in consequence , this stage still remains poorly understood . about a decade ago , however , the regions containing the earliest stages of massive star formation were identified . galactic plane surveys revealed thousands of dark patches obscuring the bright mid - infrared background ( _ iso _ , @xcite ; _ msx _ , @xcite , @xcite ; _ spitzer _ , @xcite , @xcite ) . these dark silhouettes were associated with molecular and dust emission , indicating they consist of dense molecular gas . such objects were called infrared dark clouds ( irdcs ) . the first studies characterizing them suggested that they were cold ( @xmath525 k ) , massive ( @xmath610@xmath710@xmath8 ) , and dense ( @xmath410@xmath9 @xmath10 ) molecular clouds with high column densities ( @xmath410@xmath11 @xmath12 ) @xcite . they also correspond to the densest clouds embedded within giant molecular clouds @xcite . more recent studies on irdc clumps,1 pc and a mass @xmath610@xmath710@xmath13 . we use the term `` core '' to describe a compact , dense object within a clump with a size @xmath140.1 pc and a mass @xmath1450 . ] using dust continuum emission , show that they have typical masses of @xmath6120 and sizes @xmath60.5 pc @xcite . spectral energy distributions ( seds ) of several irdc clumps reveal dust temperatures that range between 16 and 52 k , and luminosities that range from @xmath61010@xmath15 @xcite . temperatures derived using nh@xmath3 observations are lower than dust temperatures and range from @xmath61020 k @xcite . star formation activity in irdc clumps can be inferred from high temperatures and high luminosities as well as the presence of uc regions @xcite , hot cores @xcite , embedded 24 @xmath16 m sources @xcite , molecular outflows @xcite , and maser emission @xcite . the high masses , densities , and column densities of the irdc clumps , as well as the aforementioned signatures of star formation in them , indicate that irdcs are currently forming massive stars . in addition , irdcs harbor numerous candidates for the most elusive earliest phase of high - mass star formation , the `` prestellar '' or `` starless '' phase @xcite . due to the characteristics of irdcs and the large variety of evolutionary stages that they harbor , it has been suggested that they are the natal sites of all massive stars and stellar clusters @xcite . with the aim of characterizing the different evolutionary stages of clumps found in irdcs , @xcite proposed an evolutionary sequence in which `` quiescent '' clumps evolve into `` intermediate '' , `` active '' , and `` red '' clumps . this evolutionary scheme is based on the _ spitzer_/irac 3 - 8 @xmath16 m colors and the presence or absence of _ spitzer_/mips 24 @xmath16 m point - source emission . a clump is called `` quiescent '' if it contains no ir-_spitzer _ emission ( it is ir - dark ) ; `` intermediate '' if it contains either an enhanced 4.5 @xmath16 m emission , the so - called `` green fuzzies '' ( also known as extended green objects ( egos ) ; * ? ? ? * ) , or a 24 @xmath16 m source , but not both ; `` active '' if it is associated with a green fuzzy and an embedded 24 @xmath16 m source ; and `` red '' if it is associated with bright 8 @xmath16 m emission , which likely corresponds to an region . quiescent clumps are the best candidates to be in the `` prestellar '' or `` starless '' phase , and the massive quiescent clumps are the places where it is most probable that high - mass stars are in their earliest stages of evolution . an additional category is `` blue , '' which describes objects with bright 3.6 @xmath16 m emission that are predominantly unextincted stars . although several properties of irdcs have been determined in the last few years , there is still one that remains poorly explored , their chemistry . what is the chemistry in irdc clumps ? are the evolutionary stages defined by @xcite chemically distinguishable ? presently , only a few studies have focused on this subject @xcite . in this paper , we report initial results from a program aimed at better understanding the chemical evolution of irdc clumps , carrying out a multi - line survey at 3 mm . we observed several molecular lines simultaneously , which facilitates comparison between different lines by eliminating or reducing observational errors arising from uncertainties in telescope pointing and calibrations . the main goals of this project are to investigate the behavior of the different molecular tracers and look for observable changes in the chemistry as a function of the evolutionary stages proposed by @xcite . lcccccc & & 85.926260 & 20.68 & @xmath17 & if3 & 0.044 + & & 86.093983 & 19.31 & @xmath18 & if3 & 0.044 + & & 86.338735 & 4.14 & @xmath19 & if3 & 0.044 + & & 86.340167 & 4.14 & @xmath19 & if3 & 0.044 + & & 86.342256 & 4.14 & @xmath19 & if3 & 0.044 + & & 86.754330 & 4.16 & @xmath18 & if3 & 0.044 + & & 86.846998 & 6.25 & @xmath19 & if3 & 0.044 + & & 87.090859 & 4.18 & @xmath20 & if3 & 0.044 + & & 87.316925 & 4.19 & @xmath18 & if3 & 0.044 + & & 87.328624 & 4.19 & @xmath18 & if3 & 0.044 + & & 87.402004 & 4.19 & @xmath18 & if3 & 0.044 + & & 87.407165 & 4.19 & @xmath18 & if3 & 0.044 + & & 87.925252 & 10.55 & @xmath21 & if2 & 0.048 + & & 88.630416 & 4.25 & @xmath22 & if2 & 0.048 + & & 88.631847 & 4.25 & @xmath22 & if2 & 0.048 + & & 88.633936 & 4.25 & @xmath22 & if2 & 0.048 + & & 89.188526 & 4.28 & @xmath18 & if2 & 0.048 + & & 90.663574 & 4.35 & @xmath20 & if1 & 0.042 + & & 90.978989 & 24.01 & @xmath23 & if1 & 0.042 + & & 91.985316 & 20.39 & @xmath24 & if0 & 0.042 + & & 91.987089 & 13.24 & @xmath23 & if0 & 0.042 + & & 92.494303 & 6.00 & @xmath20 & if0 & 0.042 + & & 93.171913 & 4.47 & @xmath20 & if0 & 0.042 + & & 93.173772 & 4.47 & @xmath20 & if0 & 0.042 + & & 93.176261 & 4.47 & @xmath20 & if0 & 0.042 + we made single - pointing observations of molecular lines at 3 mm using the atnf mopra 22 m telescope located in australia . the source list comprises 37 irdcs , containing a total of 159 clumps . this is a sub - sample of the 38 irdcs and 190 clumps identified by @xcite using 1.2 mm dust continuum emission at 11 angular resolution . the threshold limit for a molecular line to be considered as a detection was defined at the 3@xmath25 level , where @xmath25 is the rms noise in antenna temperature ( @xmath26 ; see table [ tbl - obsparam ] ) . the clumps g015.31 mm4 , g027.75 mm5 , g028.37 mm5 , and g030.57 mm4 present no molecular line emission above the 3@xmath25 detection level ; consequently , they are not included in any analysis throughout this work . we excluded 17 clumps that show signs of star formation at ir wavelengths ( _ spitzer _ ) and are located at a different distance than the rest of the clumps within an irdc ( see section [ kinematic ] for more details ) . these clumps belong to a different molecular cloud that is situated in the line of sight . since the mopra telescope beam size ( 38 ) is sometimes comparable to the separation between clumps within the same irdc , @xmath630% of the sources are contaminated by emission from another adjacent clump . to ensure that the observed emission is attributable solely to a single clump , and not a neighboring clump , each source with an angular separation smaller than the mopra beam size from its nearest neighbor was excluded . furthermore , blue clumps are not included in any analysis because they are likely unextincted stars . therefore , this work will focus on the remaining 92 irdc clumps unless stated otherwise . @xcite and @xcite classified the initial sample of 190 irdc clumps using a combination of _ spitzer_/irac colors and the presence or absence of compact 24 @xmath16 m emission . their classifications differ in the method of estimating the 24 @xmath16 m flux measurements . @xcite used aperture photometry , and @xcite fitted a gaussian to the radial profile of the emission . due to this difference , in a few cases 24 @xmath16 m emission is detected for a particular source using one method , but not the other . in this work , to avoid ambiguities , if a clump presents 24 @xmath16 m emission from either of the two methods , a 24 @xmath16 m compact source will be assigned to the clump . table [ tbl - coor ] summarizes information for all 159 irdc clumps : names , coordinates , velocities , distances , dust temperatures , _ spitzer _ classification , and if the source is an irdc and if it is used in the present work . the observations were carried out during 2008 july august and 2010 september . they were performed with the 3 mm monolithic microwave integrated circuits ( mmic ) receiver and the mopra spectrometer ( mops ) in the `` broad - band '' mode , resulting in a total bandwidth of 8.3 ghz split over four overlapping sub - bands of 2.2 ghz . each sub - band has @xmath27 channels ( @xmath28 per each polarization ) with a channel resolution of 0.27 mhz . the velocity resolution was @xmath60.90 per channel . both polarizations were observed and averaged to improve the signal to noise ratio . the observed frequencies range between 85.540 and 93.840 ghz . this frequency range allowed us to observe 15 molecular lines simultaneously ( see table [ tbl - obsparam ] ) . the system temperatures ranged between 165 and 265 k. typical rms noise values ( @xmath26 ) for each molecular line are shown in table [ tbl - obsparam ] . the on - source integration time per object was @xmath63 minutes . the angular resolution of mopra telescope is 38 and its main beam efficiency is 0.5 at 90 ghz . all the observations were performed in position - switching mode with the off - position shifted from the target source by 1 in galactic latitude away from the galactic midplane . the telescope pointing was checked by observing nearby sio masers every @xmath61 hr , and was maintained to be better than 5 . the uncertainty in the line intensity varies with the frequency from 9% at 88 ghz to 25% at 93 ghz ( jonathan foster , private communication ) . the initial spectral processing was done using the atnf spectral analysis package ( asap ) software . the observed molecular lines , transitions , rest frequencies , upper energy levels , critical densities , sub - bands , and typical @xmath26 are listed in table [ tbl - obsparam ] . the @xmath26 for each source is given in table [ tbl - gaussian_123_012 ] . for each molecular transition , the line center velocity , line width and intensity of the line were determined from gaussian fits . molecular lines that show hyperfine structure ( n@xmath0h@xmath1 and c@xmath0h ) were fitted using a multi - gaussian function , with a fixed frequency separation between the transitions . the gaussian fit procedure was carried out in idl using the mpfitfun package @xcite . occasionally , two velocity components were detected in a source . we define the main component , i.e. , the component associated with the irdc clump , as the velocity component detected in the high - density tracers ( e.g. , n@xmath0h@xmath1 and/or h@xmath2co@xmath1 ) . the secondary velocity component typically corresponds to lower - density gas located in line of sight and they were not used in this work . when two velocity components have emission from high - density tracers , the brightest was used as the main component . the gaussian fit parameters of the main velocity component for all sources where emission was detected ( 155 ) are summarized in table [ tbl - gaussian_123_012 ] . no gaussian fits were carried out for two distinct velocity components with separations less than 3 or for self - absorbed profiles . figure [ detection_rates ] presents the detection rates , at the 3@xmath25 detection level , of the 10 most frequently detected molecular species toward the irdc clumps . for the n@xmath0h@xmath1 and c@xmath0h lines , we show the most often detected transitions : @xmath29 for n@xmath0h@xmath1 and @xmath30 for c@xmath0h . because the hcn transitions are blended , and in addition , exhibit self - absorbed profiles and wing emission , the multi - gaussian fit to the hyperfine structure was not reliable . the hcn detection rate was obtained by inspecting the spectrum for each source by eye , instead of comparing the intensity of the gaussian fit with the @xmath26 . the ch@xmath3cn , @xmath2cs , h@xmath2cn , so , and nh@xmath0d lines were detected in fewer than 8 clumps ; therefore , their detection rates are not presented in figure [ detection_rates ] . the molecular lines most often detected were hnc @xmath31 and n@xmath0h@xmath1 @xmath29 with 90 ( 98% ) and 89 ( 97% ) detections , respectively . on the other hand , the molecular line shown in figure [ detection_rates ] least often detected was the sio @xmath32 with 8 ( 9% ) detections . detection rates for all detected molecular transitions are given in table [ tbl - detection - rates ] . uncertainties presented in figure [ detection_rates ] and table [ tbl - detection - rates ] were determined assuming poisson noise . llllccccc g015.05 + 00.07 & mm1 & 18:17:50.4 & -15:53:38 & 25.2 & 2.8 & 23.5 & q & irdc + g015.05 + 00.07 & mm2 & 18:17:40.0 & -15:48:55 & 29.9 & 3.1 & 30.5 & r & irdc + g015.05 + 00.07 & mm3 & 18:17:42.4 & -15:47:03 & 28.4 & 3.0 & 23.7 & q & irdc + g015.05 + 00.07 & mm4 & 18:17:32.0 & -15:46:35 & 28.0 & 3.0 & 30.5 & q & irdc + g015.05 + 00.07 & mm5 & 18:17:40.2 & -15:49:47 & 29.9 & 3.1 & 31.0 & q & irdc + g015.31 - 00.16 & mm1 & 18:18:56.4 & -15:45:00 & 17.0 & 1.9 & ... & r & non - irdc + g015.31 - 00.16 & mm2 & 18:18:50.4 & -15:43:19 & 30.9 & 3.2 & 23.0 & i & irdc + g015.31 - 00.16 & mm3 & 18:18:45.3 & -15:41:58 & 31.1 & 3.2 & 23.7 & q & irdc + g015.31 - 00.16 & mm4 & 18:18:48.0 & -15:44:22 & 31.1 & 1.9 & ... & q & no detection + g015.31 - 00.16 & mm5 & 18:18:49.1 & -15:42:47 & 31.2 & 3.2 & 28.0 & q & irdc + g018.82 - 00.28 & mm1 & 18:25:56.1 & -12:42:48 & 41.7 & 3.5 & ... & i & non - irdc + g018.82 - 00.28 & mm2 & 18:26:23.4 & -12:39:37 & 63.1 & 4.7 & 38.0 & r & irdc + g018.82 - 00.28 & mm3 & 18:25:52.6 & -12:44:37 & 44.8 & 3.7 & 34.0 & a & irdc + g018.82 - 00.28 & mm4 & 18:26:15.5 & -12:41:32 & 65.5 & 4.8 & 17.0 & i & irdc + g018.82 - 00.28 & mm6 & 18:26:18.4 & -12:41:15 & 65.8 & 4.8 & 23.7 & q & irdc + g019.27 + 00.07 & mm1 & 18:25:58.5 & -12:03:59 & 26.7 & 2.4 & ... & b & bright blue + g019.27 + 00.07 & mm2 & 18:25:52.6 & -12:04:48 & 26.7 & 2.4 & 30.0 & a & irdc + g022.35 + 00.41 & mm1 & 18:30:24.4 & -09:10:34 & 52.7 & 3.9 & 20.0 & a & irdc + g022.35 + 00.41 & mm2 & 18:30:24.2 & -09:12:44 & 59.9 & 4.2 & 44.0 & r & irdc + g023.60 + 00.00 & mm1 & 18:34:11.6 & -08:19:06 & 106.5 & 7.0 & ... & a & non - irdc + g023.60 + 00.00 & mm2 & 18:34:21.1 & -08:18:07 & 53.7 & 3.9 & ... & a & blended + g023.60 + 00.00 & mm3 & 18:34:10.0 & -08:18:28 & 105.8 & 6.9 & ... & r & non - irdc + g023.60 + 00.00 & mm4 & 18:34:23.0 & -08:18:21 & 53.6 & 3.9 & ... & r & blended + g023.60 + 00.00 & mm5 & 18:34:09.5 & -08:18:00 & 104.8 & 6.8 & ... & a & non - irdc + g023.60 + 00.00 & mm7 & 18:34:21.1 & -08:17:11 & 54.0 & 3.9 & 44.0 & i & irdc + g023.60 + 00.00 & mm9 & 18:34:22.5 & -08:16:04 & 54.3 & 3.9 & 32.5 & q & irdc + g024.08 + 00.04 & mm1 & 18:34:57.0 & -07:43:26 & 114.1 & 7.8 & ... & r & non - irdc + g024.08 + 00.04 & mm2 & 18:34:51.1 & -07:45:32 & 114.1 & 7.8 & 29.0 & q & irdc + g024.08 + 00.04 & mm3 & 18:35:02.2 & -07:45:25 & 51.6 & 3.7 & 29.0 & q & irdc + g024.08 + 00.04 & mm4 & 18:35:02.6 & -07:45:56 & 52.2 & 3.7 & 30.0 & q & irdc + g024.33 + 00.11 & mm1 & 18:35:07.9 & -07:35:04 & 114.1 & 7.8 & ... & r & blended + g024.33 + 00.11 & mm2 & 18:35:34.5 & -07:37:28 & 118.2 & 7.7 & 32.5 & i & irdc + g024.33 + 00.11 & mm3 & 18:35:27.9 & -07:36:18 & 117.6 & 7.7 & 31.5 & r & irdc + g024.33 + 00.11 & mm4 & 18:35:19.4 & -07:37:17 & 115.0 & 7.8 & 31.0 & q & irdc + g024.33 + 00.11 & mm5 & 18:35:33.8 & -07:36:42 & 117.2 & 7.7 & 32.0 & i & irdc + g024.33 + 00.11 & mm6 & 18:35:07.7 & -07:34:33 & 114.3 & 7.7 & ... & i & blended + g024.33 + 00.11 & mm7 & 18:35:09.8 & -07:39:48 & 99.6 & 6.3 & 23.7 & q & irdc + g024.33 + 00.11 & mm8 & 18:35:23.4 & -07:37:21 & 113.8 & 7.8 & 31.5 & q & irdc + g024.33 + 00.11 & mm9 & 18:35:26.5 & -07:36:56 & 119.2 & 7.7 & 52.0 & r & irdc + g024.33 + 00.11 & mm11 & 18:35:05.1 & -07:35:58 & 113.3 & 7.8 & 26.0 & q & irdc + g024.60 + 00.08 & mm1 & 18:35:41.1 & -07:18:30 & 53.4 & 3.8 & ... & a & blended + g024.60 + 00.08 & mm2 & 18:35:39.3 & -07:18:51 & 115.2 & 7.7 & 23.0 & i & irdc + g024.60 + 00.08 & mm3 & 18:35:40.2 & -07:18:37 & 53.8 & 3.8 & ... & i & blended + g024.60 + 00.08 & mm4 & 18:35:35.7 & -07:18:09 & 53.2 & 3.8 & ... & q & blended + g025.04 - 00.20 & mm1 & 18:38:10.2 & -07:02:34 & 64.0 & 4.3 & ... & a & blended + g025.04 - 00.20 & mm2 & 18:38:17.7 & -07:02:51 & 63.5 & 4.3 & 28.0 & i & irdc + g025.04 - 00.20 & mm3 & 18:38:10.2 & -07:02:44 & 63.7 & 4.3 & ... & q & blended + g025.04 - 00.20 & mm4 & 18:38:13.7 & -07:03:12 & 63.8 & 4.3 & 29.5 & i & irdc + g025.04 - 00.20 & mm5 & 18:38:12.0 & -07:02:44 & 63.9 & 4.3 & ... & q & blended + g027.75 + 00.16 & mm1 & 18:41:19.9 & -04:32:20 & 78.3 & 5.0 & ... & a & blended + g027.75 + 00.16 & mm2 & 18:41:33.0 & -04:33:44 & 51.4 & 3.5 & 24.0 & q & irdc + g027.75 + 00.16 & mm3 & 18:41:16.8 & -04:31:55 & 78.7 & 5.0 & ... & i & blended + g027.75 + 00.16 & mm5 & 18:41:23.6 & -04:30:42 & 79.1 & 4.8 & ... & q & no detection + g027.94 - 00.47 & mm1 & 18:44:03.6 & -04:38:00 & 45.4 & 3.1 & 44.0 & r & irdc + g027.97 - 00.42 & mm1 & 18:43:52.8 & -04:36:13 & 44.6 & 3.1 & ... & a & blended + g027.97 - 00.42 & mm2 & 18:43:58.0 & -04:34:24 & 19.9 & 1.5 & ... & r & non - irdc + g027.97 - 00.42 & mm3 & 18:43:54.9 & -04:36:08 & 45.9 & 3.2 & ... & q & blended + g028.04 - 00.46 & mm1 & 18:44:08.5 & -04:33:22 & 45.8 & 3.2 & 38.0 & a & irdc + g028.08 + 00.07 & mm1 & 18:42:20.3 & -04:16:42 & 81.4 & 5.2 & 22.0 & i & irdc + g028.10 - 00.45 & mm1 & 18:44:12.9 & -04:29:45 & 47.1 & 3.2 & ... & q & blended + g028.10 - 00.45 & mm2 & 18:44:14.3 & -04:29:48 & 46.9 & 3.2 & ... & q & blended + g028.23 - 00.19 & mm1 & 18:43:30.7 & -04:13:12 & 80.0 & 5.1 & 22.0 & q & irdc + g028.23 - 00.19 & mm2 & 18:43:29.0 & -04:12:16 & 80.8 & 5.1 & ... & i & blended + g028.23 - 00.19 & mm3 & 18:43:30.0 & -04:12:33 & 80.2 & 5.1 & ... & q & blended + g028.28 - 00.34 & mm1 & 18:44:15.0 & -04:17:54 & 48.7 & 3.3 & ... & r & blended + g028.28 - 00.34 & mm2 & 18:44:21.3 & -04:17:37 & 84.9 & 5.6 & ... & r & non - irdc + g028.28 - 00.34 & mm3 & 18:44:13.4 & -04:18:05 & 49.7 & 3.4 & ... & r & blended + g028.28 - 00.34 & mm4 & 18:44:11.4 & -04:17:22 & 48.8 & 3.4 & 34.5 & a & irdc + g028.37 + 00.07 & mm1 & 18:42:52.1 & -03:59:45 & 78.1 & 5.0 & 33.0 & a & irdc + g028.37 + 00.07 & mm2 & 18:42:37.6 & -04:02:05 & 80.8 & 5.1 & 39.0 & i & irdc + g028.37 + 00.07 & mm3 & 18:43:03.1 & -04:06:24 & 99.8 & 6.4 & ... & r & non - irdc + g028.37 + 00.07 & mm4 & 18:42:50.7 & -04:03:15 & 79.3 & 5.0 & 32.0 & a & irdc + g028.37 + 00.07 & mm5 & 18:42:26.8 & -04:01:30 & 78.6 & 4.8 & ... & r & no detection + g028.37 + 00.07 & mm6 & 18:42:49.0 & -04:02:23 & 80.0 & 5.1 & 23.0 & a & irdc + g028.37 + 00.07 & mm7 & 18:42:56.3 & -04:07:31 & 46.6 & 3.2 & ... & r & non - irdc + g028.37 + 00.07 & mm8 & 18:42:49.7 & -04:09:54 & 107.2 & 7.5 & ... & r & non - irdc + g028.37 + 00.07 & mm9 & 18:42:46.7 & -04:04:08 & 79.4 & 5.0 & 24.5 & q & irdc + g028.37 + 00.07 & mm10 & 18:42:54.0 & -04:02:30 & 79.1 & 5.0 & ... & a & blended + g028.37 + 00.07 & mm11 & 18:42:42.7 & -04:01:44 & 80.9 & 5.1 & 36.0 & a & irdc + g028.37 + 00.07 & mm12 & 18:43:09.9 & -04:06:52 & 34.8 & 2.5 & 23.7 & q & irdc + g028.37 + 00.07 & mm13 & 18:42:41.8 & -03:57:08 & 77.5 & 4.9 & 30.4 & i & irdc + g028.53 - 00.25 & mm1 & 18:44:18.0 & -03:59:34 & 86.8 & 5.5 & ... & q & blended + g028.53 - 00.25 & mm2 & 18:44:15.7 & -03:59:41 & 85.6 & 5.4 & ... & a & blended + g028.53 - 00.25 & mm3 & 18:44:16.0 & -04:00:48 & 86.4 & 5.5 & 22.0 & q & irdc + g028.53 - 00.25 & mm4 & 18:44:18.6 & -04:00:05 & 86.3 & 5.4 & ... & i & blended + g028.53 - 00.25 & mm5 & 18:44:17.0 & -04:02:04 & 87.0 & 5.7 & 30.0 & i & irdc + g028.53 - 00.25 & mm6 & 18:44:17.8 & -04:00:05 & 86.4 & 5.5 & ... & q & blended + g028.53 - 00.25 & mm7 & 18:44:23.7 & -04:02:09 & 88.6 & 5.6 & 23.0 & q & irdc + g028.53 - 00.25 & mm8 & 18:44:22.0 & -04:01:35 & 88.6 & 5.6 & 28.0 & q & irdc + g028.53 - 00.25 & mm9 & 18:44:19.3 & -03:58:05 & 87.1 & 5.7 & 28.0 & i & irdc + g028.53 - 00.25 & mm10 & 18:44:18.5 & -03:58:43 & 86.8 & 5.5 & 30.0 & q & irdc + g028.67 + 00.13 & mm1 & 18:43:03.1 & -03:41:41 & 84.3 & 5.3 & 40.0 & r & irdc + g028.67 + 00.13 & mm2 & 18:43:07.1 & -03:44:01 & 79.2 & 5.0 & 23.0 & i & irdc + g028.67 + 00.13 & mm3 & 18:42:58.2 & -03:48:20 & 103.9 & 6.9 & ... & a & non - irdc + g030.14 - 00.06 & mm1 & 18:46:35.7 & -02:31:03 & 86.8 & 5.5 & 27.0 & a & irdc + g030.57 - 00.23 & mm1 & 18:48:00.0 & -02:07:20 & 90.7 & 5.8 & 34.5 & a & irdc + g030.57 - 00.23 & mm2 & 18:47:58.7 & -02:15:20 & 111.3 & 7.3 & ... & r & non - irdc + g030.57 - 00.23 & mm3 & 18:47:54.5 & -02:11:15 & 95.9 & 6.3 & 23.0 & i & irdc + g030.57 - 00.23 & mm4 & 18:48:01.8 & -02:12:35 & 86.2 & 5.2 & ... & q & no detection + g030.97 - 00.14 & mm1 & 18:48:21.6 & -01:48:27 & 77.8 & 5.0 & 37.0 & a & irdc + g031.02 - 00.10 & mm1 & 18:48:09.9 & -01:45:17 & 76.9 & 4.9 & ... & b & bright blue + g031.97 + 00.07 & mm1 & 18:49:36.3 & -00:45:45 & 95.0 & 6.5 & ... & a & blended + g031.97 + 00.07 & mm2 & 18:49:36.0 & -00:46:16 & 94.6 & 6.5 & ... & q & blended + g031.97 + 00.07 & mm3 & 18:49:32.3 & -00:47:02 & 94.1 & 6.4 & ... & q & blended + g031.97 + 00.07 & mm4 & 18:49:33.0 & -00:47:33 & 95.5 & 6.6 & ... & i & blended + g031.97 + 00.07 & mm5 & 18:49:21.9 & -00:50:35 & 96.5 & 6.8 & 35.0 & i & irdc + g031.97 + 00.07 & mm6 & 18:49:35.0 & -00:46:44 & 94.8 & 6.5 & ... & q & blended + g031.97 + 00.07 & mm7 & 18:49:28.4 & -00:48:54 & 93.9 & 6.4 & 27.0 & q & irdc + g031.97 + 00.07 & mm8 & 18:49:29.1 & -00:48:12 & 94.5 & 6.4 & 47.0 & a & irdc + g033.69 - 00.01 & mm1 & 18:52:58.8 & 00:42:37 & 105.7 & 7.1 & 25.0 & r & irdc + g033.69 - 00.01 & mm2 & 18:52:49.9 & 00:37:57 & 104.8 & 7.1 & 41.0 & r & irdc + g033.69 - 00.01 & mm3 & 18:52:50.8 & 00:36:43 & 103.1 & 7.1 & 43.0 & r & irdc + g033.69 - 00.01 & mm4 & 18:52:56.4 & 00:43:08 & 106.2 & 7.1 & 28.0 & a & irdc + g033.69 - 00.01 & mm5 & 18:52:47.8 & 00:36:47 & 105.5 & 7.1 & 37.0 & a & irdc + g033.69 - 00.01 & mm11 & 18:52:56.2 & 00:41:48 & 107.0 & 7.1 & 25.0 & q & irdc + g034.43 + 00.24 & mm1 & 18:53:18.0 & 01:25:24 & 57.9 & 3.8 & 38.0 & a & irdc + g034.43 + 00.24 & mm2 & 18:53:18.6 & 01:24:40 & 57.4 & 3.8 & ... & r & blended + g034.43 + 00.24 & mm3 & 18:53:20.4 & 01:28:23 & 59.4 & 3.9 & ... & a & blended + g034.43 + 00.24 & mm4 & 18:53:19.0 & 01:24:08 & 57.6 & 3.8 & ... & a & blended + g034.43 + 00.24 & mm5 & 18:53:19.8 & 01:23:30 & 58.0 & 3.8 & 23.0 & a & irdc + g034.43 + 00.24 & mm6 & 18:53:18.6 & 01:27:48 & 58.6 & 3.8 & ... & i & blended + g034.43 + 00.24 & mm7 & 18:53:18.3 & 01:27:13 & 58.1 & 3.8 & 29.5 & i & irdc + g034.43 + 00.24 & mm8 & 18:53:16.4 & 01:26:20 & 57.2 & 3.7 & 43.0 & a & irdc + g034.43 + 00.24 & mm9 & 18:53:18.4 & 01:28:14 & 58.7 & 3.8 & ... & q & blended + g034.77 - 00.55 & mm1 & 18:56:48.2 & 01:18:47 & 44.2 & 2.9 & 40.4 & r & irdc + g034.77 - 00.55 & mm2 & 18:56:50.3 & 01:23:16 & 42.0 & 2.8 & ... & q & blended + g034.77 - 00.55 & mm3 & 18:56:44.7 & 01:20:42 & 43.1 & 2.8 & 30.4 & i & irdc + g034.77 - 00.55 & mm4 & 18:56:48.9 & 01:23:34 & 42.1 & 2.8 & ... & q & blended + g035.39 - 00.33 & mm1 & 18:56:41.2 & 02:09:52 & 64.3 & 4.2 & ... & r & non - irdc + g035.39 - 00.33 & mm2 & 18:56:59.2 & 02:04:53 & 54.0 & 3.5 & ... & a & non - irdc + g035.39 - 00.33 & mm3 & 18:57:05.3 & 02:06:29 & 55.3 & 3.6 & ... & r & non - irdc + g035.39 - 00.33 & mm4 & 18:57:06.7 & 02:08:23 & 45.3 & 3.0 & ... & a & blended + g035.39 - 00.33 & mm5 & 18:57:08.8 & 02:08:09 & 90.1 & 6.9 & ... & q & blended + g035.39 - 00.33 & mm6 & 18:57:08.4 & 02:09:09 & 45.5 & 3.0 & ... & a & blended + g035.39 - 00.33 & mm7 & 18:57:08.1 & 02:10:50 & 45.9 & 3.0 & 34.0 & a & irdc + g035.59 - 00.24 & mm1 & 18:57:02.3 & 02:17:04 & 50.4 & 3.3 & 36.0 & r & irdc + g035.59 - 00.24 & mm2 & 18:57:07.4 & 02:16:14 & 45.3 & 3.0 & 24.0 & a & irdc + g035.59 - 00.24 & mm3 & 18:57:11.6 & 02:16:08 & 45.0 & 3.0 & 30.5 & a & irdc + g036.67 - 00.11 & mm1 & 18:58:39.6 & 03:16:16 & 53.5 & 3.5 & 20.0 & q & irdc + g036.67 - 00.11 & mm2 & 18:58:35.6 & 03:15:06 & 53.3 & 3.5 & 16.0 & q & irdc + g038.95 - 00.47 & mm1 & 19:04:07.4 & 05:08:48 & 42.4 & 2.8 & 16.0 & i & irdc + g038.95 - 00.47 & mm2 & 19:04:03.4 & 05:07:56 & 42.1 & 2.8 & 40.0 & r & irdc + g038.95 - 00.47 & mm3 & 19:04:07.4 & 05:09:44 & 42.4 & 2.8 & 28.0 & i & irdc + g038.95 - 00.47 & mm4 & 19:04:00.6 & 05:09:06 & 42.1 & 2.8 & 43.0 & a & irdc + g048.65 - 00.29 & mm2 & 19:21:47.6 & 13:49:22 & 33.8 & 2.5 & ... & q & blended + g053.11 + 00.05 & mm1 & 19:29:17.2 & 17:56:21 & 21.8 & 1.8 & 38.0 & r & irdc + g053.11 + 00.05 & mm2 & 19:29:20.2 & 17:57:06 & 22.9 & 1.9 & 36.0 & a & irdc + g053.11 + 00.05 & mm3 & 19:29:00.6 & 17:55:11 & 2.6 & 0.5 & ... & r & non - irdc + g053.11 + 00.05 & mm4 & 19:29:20.4 & 17:55:04 & 21.7 & 1.8 & 40.0 & a & irdc + g053.11 + 00.05 & mm5 & 19:29:26.3 & 17:54:53 & 21.7 & 1.8 & 45.0 & r & irdc + g053.25 + 00.04 & mm1 & 19:29:39.0 & 18:01:42 & 24.9 & 2.0 & 35.0 & i & irdc + g053.25 + 00.04 & mm2 & 19:29:33.0 & 18:01:00 & 24.3 & 2.0 & ... & b & bright blue + g053.25 + 00.04 & mm3 & 19:29:44.0 & 17:58:47 & 23.5 & 1.9 & 35.0 & i & irdc + g053.25 + 00.04 & mm4 & 19:29:34.5 & 18:01:39 & 24.5 & 2.0 & 37.0 & a & irdc + g053.25 + 00.04 & mm5 & 19:29:39.4 & 17:58:40 & 23.7 & 1.9 & 35.0 & i & irdc + g053.25 + 00.04 & mm6 & 19:29:31.5 & 17:59:50 & 23.8 & 1.9 & 46.0 & a & irdc + g053.31 + 00.00 & mm1 & 19:29:50.0 & 18:05:07 & 22.0 & 1.8 & ... & i & blended + g053.31 + 00.00 & mm2 & 19:29:42.1 & 18:03:57 & 25.2 & 2.0 & 25.0 & q & irdc + g053.31 + 00.00 & mm3 & 19:29:49.7 & 18:04:39 & 22.2 & 1.8 & ... & i & blended + h@xmath1 and c@xmath0h detection rates correspond to those of the @xmath33 and @xmath30 transitions , respectively . error bars are the poisson noise ( i.e. , the root square of the number of detections ) . values on bars are number of detections for a given molecule . ] figures [ fig - quiescent ] , [ fig - intermediate ] , [ fig - active ] , and [ fig - red ] show representative examples of spectra for one clump of each proposed evolutionary stage : the quiescent g028.53 mm3 , the intermediate g025.04 mm4 , the active g034.43 mm1 , and the red g034.77 mm1 , respectively . the _ spitzer_/irac image of the corresponding host irdc is also displayed in each figure . the active clump g034.43 mm1 ( figure [ fig - active ] ) exhibits more molecular line emission than the quiescent clump g028.53 mm3 ( figure [ fig - quiescent ] ) . this pattern is typical through the sample . it is expected that when star - forming clumps evolve to later stages , they should display a rich molecular spectrum with numerous lines because the proto - stars heat their environment releasing more complex molecules from dust mantles into the gas phase ; in contrast , in earlier stages clumps should show a sparse spectrum with a smaller number of detectable molecular lines because many molecules are frozen out onto grain surfaces . however , one must be aware that although the analysis of detection rates may reflect the chemical composition of clumps to some extent , it does not give unambiguous conclusions because detection rates may just show the dependence on h@xmath0 column density . a more definitive conclusion must be drawn from molecular abundances ( see sections [ chemistry ] and [ molecules ] ) . figure [ number_lines ] presents histograms which compare the number of detected molecular transitions for each group . more molecular transitions are detected toward clumps that show signs of star formation . the median number of detections for quiescent clumps is 6 molecular transitions . on the other hand , intermediate and active clumps have a median of 7 and 9 detected molecular transitions , respectively . to quantify the differences between quiescent and active clumps , we have calculated the probability that these two distributions in figure [ number_lines ] arise from the same underlying population using the kolmogorov smirnov ( k s ) test . s test is a more robust method for measuring the similarity between two distributions than the comparison of the median and/or mean values . through this paper , we will mention that the difference between two distributions is statistically significant when the k - s test gives a percentage lower than 5% . the probability that the quiescent and active clumps are derived from the same parent population is 0.1% . llcccc g015.05 + 00.07 & mm1 & 0.04 & 0.38 @xmath34 0.03 & 25.21 @xmath34 0.10 & 3.18 @xmath34 0.31 + g015.05 + 00.07 & mm2 & 0.04 & 0.19 @xmath34 0.03 & 29.87 @xmath34 0.20 & 3.64 @xmath34 0.62 + g015.05 + 00.07 & mm3 & 0.04 & 0.18 @xmath34 0.03 & 28.41 @xmath34 0.21 & 2.50 @xmath34 0.54 + g015.05 + 00.07 & mm4 & 0.04 & 0.44 @xmath34 0.03 & 27.95 @xmath34 0.08 & 2.37 @xmath34 0.21 + g015.05 + 00.07 & mm5 & 0.05 & 0.25 @xmath34 0.05 & 29.93 @xmath34 0.16 & 2.03 @xmath34 0.46 + the median for red clumps is 6 detections . red clumps ( which show bright 8 @xmath16 m emission ) likely correspond to embedded hii regions , a more evolved stage than active clumps . despite this , they have the same median value as quiescent clumps . in the histogram , red clumps seem to have a bimodal distribution with a first population of 8 clumps showing a low number of detected molecular transitions ( left side of the histogram ) and a second one of 7 clumps presenting a higher number ( right side of the histogram ) . moreover , we note that the first population only has one maser detection , and the second population has five @xcite . these two different populations for red clumps may be the result of the evolutionary state of the hii region . we expect to detect more molecular lines from an early hii region compared to a late hii region because , since the early hii region has not had as much time to ionize the surrounding gas , it should have more molecular gas near the central heating star(s ) . the molecular spectrum of a late stage hii region , where much of the gas has been ionized , should be more sparse . h@xmath1 and c@xmath0h detection rates correspond to the @xmath33 and @xmath30transitions , respectively . the evolutionary sequence name and the total number of sources are given on the top right corner of each panel . hnc and n@xmath0h@xmath1 lines are detected in almost every irdc clumps at every evolutionary stage . on the other hand , hc@xmath3n , hnco , and sio lines are predominantly detected in later stages of evolution . error bars are the poisson noise ( i.e. , the root square of the number of detections ) . ] the active clump g034.43 mm1 ( figure [ fig - active ] ) exhibits a spectrum of a `` hot core . '' it shows emission in the hnco , hc@xmath3n , and ch@xmath3cn lines , which require high densities and temperatures for excitation ( @xmath35 @xmath10 and @xmath36 k ) , and depend on the release of their parent molecules from dust grains to be formed . in addition , it shows sio emission , which is usually associated with molecular outflows since sio abundance is highly enhanced by shocks ( e.g. , * ? ? ? * ; * ? ? ? * ) . in contrast , the quiescent clump g028.53 mm3 ( figure [ fig - quiescent ] ) presents no emission of hnco , hc@xmath3n , ch@xmath3cn , and sio over the @xmath37 limit . the intermediate clump g025.04 mm4 ( figure [ fig - intermediate ] ) seems to be in an intermediate chemical state , only exhibiting one high - excitation line ( hnco ) . to carry out a more comprehensive study of the full sample of irdc clumps , we have split the histogram of detection rates ( figure [ detection_rates ] ) into four new histograms . in figure [ detection_rates_per_category ] the detection rates for each proposed evolutionary sequence are presented . the detection rates for the quiescent clumps are displayed in every panel for comparison . quiescent , intermediate , active , and red clumps have 28 , 23 , 26 , and 15 members , respectively . from the histograms in figure [ detection_rates_per_category ] , it is seen that the hnc @xmath31 , n@xmath0h@xmath1 @xmath29 , and hco@xmath1 @xmath31 lines are present in almost every irdc clump at any evolutionary stage . red and active clumps have a high detection rate ( @xmath695% ) for the hcn @xmath31 line , comparable with the three lines previously mentioned , while in less evolved clumps this value is lower ( @xmath669% ) . there is a high enhancement in the detection of the brightest c@xmath0h line , transition @xmath30 , toward active and intermediate clumps with respect to quiescent clumps . the percentages of detection for the h@xmath2co@xmath1 @xmath31 line are roughly the same for red , intermediate , and quiescent clumps ; however , the value for active clumps increases by @xmath630% with respect to the other three evolutionary stages . the detection rates of the hn@xmath2c @xmath31 line show no significant difference between red and intermediate clumps with respect to quiescent clumps ; on the other hand , the value for active clumps increases by @xmath630% . the most remarkable difference between detection rates for the different evolutionary stages is found for the hc@xmath3n @xmath38 line . the detection of this line is significantly high for all stages with ir signs of star formation . hc@xmath3n is only detected in one quiescent source , and the difference between the detection rate in active and quiescent clumps is @xmath640% . while the hnco @xmath39 line is only detected twice ( 7% ) in quiescent clumps , active and intermediate clumps have a hnco @xmath39 detection rate of @xmath627% , indicating that this line is also observed more often toward clumps with current star formation . in the quiescent , intermediate , and red clumps , sio @xmath40 emission was rarely detected . in active clumps , the detection rate of sio reaches the 19% . the widths of molecular lines give information about the turbulence of clumps . although we can not rule out that clumps have organized velocity structures , such as rotation , in this paper we will assume that turbulence dominates the widths of the lines . to investigate if turbulence is different in clumps with and without signs of star formation , we made histograms of the number distributions of full width at half maximum line widths ( @xmath41v ) for the four evolutionary stages using the three most commonly detected lines : n@xmath0h@xmath1 , hco@xmath1 , and hnc . in figure [ fwhm - histo ] , the number distributions of n@xmath0h@xmath1 @xmath41v for each evolutionary sequence is presented . the range of line widths varies between 1.6 and 4.6 for the full sample of irdc clumps . the median values for quiescent , intermediate , active , and red clumps are 2.7 , 2.8 , 3.0 , and 3.4 , respectively . using the k s test , we calculated a probability of 67% that active and quiescent , and 6% that red and quiescent distributions originate from the same parent populations . although the k s test shows that there is little difference between the active and quiescent clumps , it seems that there is an increase in @xmath41v from clumps with no apparent star formation to active and red clumps . we think this trend is less clearly shown by the k s test for active and quiescent distributions because of the small number of members of each group ; in spite of this , the k s test shows a difference between the red and quiescent populations because the shape of the number distributions are more dissimilar . we conclude that the slightly broader widths found in the n@xmath0h@xmath1 line toward red and active clumps , compared with quiescent and intermediate clumps , is produced by the rise in the turbulence from the ongoing star formation activity . the number distributions of hco@xmath1 and hnc @xmath41v show no clear trend , and their histograms are not presented . the range of line widths varies between 2.0 and 13.0 for hco@xmath1 , and between 1.0 and 7.6 for hnc . the @xmath41v median values of hco@xmath1 for quiescent , intermediate , active , and red clumps are 3.8 , 4.4 , 3.6 , and 4.0 ; and for hnc are 3.2 , 3.8 , 3.4 , and 2.8 , respectively . we identify three reasons that may explain why hco@xmath1 and hnc do not present a similar trend to that one shown by n@xmath0h@xmath1 . first , the hco@xmath1 and hnc lines are optically thick ( see section [ opacity ] ) and consequently , the line widths can be broadened by opacity ( e.g. , * ? ? ? second , the hco@xmath1 line can show broadening and wing emission produced by molecular outflow activity . certainly , this is what is happening in g018.82 mm3 , which has a @xmath41v of 13.0 in the hco@xmath1 line . finally , because the mopra beam is @xmath638 ( 0.8 pc at the average distance of 4.3 kpc ) , we are observing the bulk motions of the gas and some level of clumping may also be present . this also could explain why the trend of broader n@xmath0h@xmath1 lines toward more evolved objects is not as clear as we expect . however , since n@xmath0h@xmath1 is thought to be a higher density tracer ( e.g. , * ? ? ? * ; * ? ? ? * ) , its emission should emanate mostly from the center of the clumps and thus we still can see a weak trend . lclllll & & 27 @xmath34 5 ( 96% ) & 22 @xmath34 5 ( 96% ) & 26 @xmath34 5 ( 100% ) & 15 @xmath34 4 ( 100% ) & 90 @xmath34 10 ( 98% ) + & & 22 @xmath34 5 ( 79% ) & 22 @xmath34 5 ( 96% ) & 25 @xmath34 5 ( 96% ) & 9 @xmath34 3 ( 60% ) & 78 @xmath34 9 ( 85% ) + & & 27 @xmath34 5 ( 96% ) & 23 @xmath34 5 ( 100% ) & 26 @xmath34 5 ( 100% ) & 13 @xmath34 4 ( 87% ) & 89 @xmath34 9 ( 97% ) + & & 8 @xmath34 3 ( 29% ) & 12 @xmath34 4 ( 52% ) & 17 @xmath34 4 ( 65% ) & 5 @xmath34 2 ( 33% ) & 42 @xmath34 7 ( 46% ) + & & 25 @xmath34 5 ( 89% ) & 17 @xmath34 4 ( 74% ) & 25@xmath34 5 ( 96% ) & 14 @xmath34 4 ( 93% ) & 81 @xmath34 9 ( 88% ) + & & 19 @xmath34 4 ( 68% ) & 16 @xmath34 4 ( 70% ) & 25@xmath34 5 ( 96% ) & 14 @xmath34 4 ( 93% ) & 74 @xmath34 9 ( 80% ) + & & 9 @xmath34 3 ( 32% ) & 15 @xmath34 4 ( 65% ) & 19 @xmath34 4 ( 73% ) & 8 @xmath34 3 ( 53% ) & 51 @xmath34 7 ( 55% ) + & & 1 @xmath34 1 ( 4% ) & 3 @xmath34 2 ( 13% ) & 10 @xmath34 3 ( 39% ) & 4 @xmath34 2 ( 27% ) & 18 @xmath34 4 ( 20% ) + & & 1 @xmath34 1 ( 4% ) & 6 @xmath34 3 ( 26% ) & 9 @xmath34 3 ( 35% ) & 3 @xmath34 2 ( 20% ) & 19 @xmath34 4 ( 21% ) + & & 1 @xmath34 1 ( 4% ) & 1 @xmath34 1 ( 4% ) & 4 @xmath34 2 ( 15% ) & 2 @xmath34 1 ( 13% ) & 8 @xmath34 3 ( 9% ) + & & 13 @xmath34 4 ( 46% ) & 11 @xmath34 3 ( 48% ) & 20 @xmath34 5 ( 77% ) & 6 @xmath34 3 ( 40% ) & 50 @xmath34 7 ( 54% ) + & & 7 @xmath34 3 ( 25% ) & 8 @xmath34 3 ( 35% ) & 14 @xmath34 4 ( 54% ) & 5 @xmath34 2 ( 33% ) & 34 @xmath34 6 ( 37% ) + & & 1 @xmath34 1 ( 4% ) & 4 @xmath34 2 ( 17% ) & 12 @xmath34 4 ( 46% ) & 5 @xmath34 2 ( 33% ) & 22 @xmath34 5 ( 24% ) + & & 2 @xmath34 1 ( 7% ) & 6 @xmath34 3 ( 26% ) & 7 @xmath34 3 ( 27% ) & 3 @xmath34 2 ( 20% ) & 18 @xmath34 4 ( 20% ) + & & 1 @xmath34 1 ( 4% ) & 0 @xmath34 0 ( 0% ) & 5 @xmath34 2 ( 19% ) & 2 @xmath34 1 ( 13% ) & 8 @xmath34 3 ( 9% ) + & & 0 @xmath34 0 ( 0% ) & 1 @xmath34 1 ( 4% ) & 3 @xmath34 2 ( 12% ) & 2 @xmath34 1 ( 13% ) & 6 @xmath34 2 ( 7% ) + & & 0 @xmath34 0 ( 0% ) & 0 @xmath34 0 ( 0% ) & 3 @xmath34 2 ( 12% ) & 0 @xmath34 0 ( 0% ) & 3 @xmath34 2 ( 3% ) + & & 0 @xmath34 0 ( 0% ) & 0 @xmath34 0 ( 0% ) & 0 @xmath34 0 ( 0% ) & 1 @xmath34 1 ( 7% ) & 1 @xmath34 1 ( 1% ) + & & 0 @xmath34 0 ( 0% ) & 0 @xmath34 0 ( 0% ) & 2 @xmath34 1 ( 8% ) & 0 @xmath34 0 ( 0% ) & 2 @xmath34 1 ( 2% ) + & @xmath42 & 0 @xmath34 0 ( 0% ) & 0 @xmath34 0 ( 0% ) & 1 @xmath34 1 ( 4% ) & 0 @xmath34 0 ( 0% ) & 1 @xmath34 1 ( 1% ) + h@xmath1 linewidths ( @xmath41v ) for each evolutionary sequence . the name of the evolutionary stage is given on the top left corner of each panel . the vertical dashed lines indicate the median values of the @xmath41v for each distribution . although the k - s test shows that active and quiescent clumps probably originate from indistinguishable parent populations , the k - s test shows that red and quiescent distributions are different . the median values of the @xmath41v increase with the evolution of the clumps . ] variations in n@xmath0h@xmath1 line widths for different evolutionary stages toward irdc clumps have also been observed previously by @xcite and recently by @xcite . @xcite observed 22 _ msx _ dark objects and 7 _ msx _ sources located at distances less than 4.5 kpc . _ msx _ dark objects are defined by @xcite as objects with 8 @xmath16 m extinction features , although some of them have _ spitzer _ 24 @xmath16 m pointlike sources ; thus , _ msx _ dark objects correspond to our intermediate and quiescent clumps . the _ msx _ sources correspond to our red clumps . @xcite found that the @xmath41v of n@xmath0h@xmath1 lines range between 0.6 and 4.0 , and the mean values for _ msx _ dark objects and _ msx _ sources are 2.3 and 2.5 , respectively . @xcite observed 13 `` quiescent '' clumps ( as defined by us ) , 11 `` middle '' clumps ( which correspond to our intermediate and active clumps ) , and 13 `` active '' clumps ( which correspond to our red clumps ) . they found that the range of @xmath41v of n@xmath0h@xmath1 lines varies between 0.6 and 2.8 ; and `` quiescent , '' `` middle , '' and `` active '' clumps have @xmath41v mean values of 1.4 , 1.7 , and 2.2 , respectively . if the assumption that turbulence increases with star formation activity producing broader molecular line widths in more evolved stages is correct , then the results on n@xmath0h@xmath1 line widths found by @xcite , @xcite , and this work support the evolutionary sequence proposed by @xcite for irdc clumps . figure [ fwhm - plot ] shows the @xmath41v of the most frequently detected molecular lines against the widths of n@xmath0h@xmath1 . the linewidths for hnc , hco@xmath1 , and sio lines show a high degree of scatter . in the case of hnc and hco@xmath1 , this scatter is likely produced by the high line opacities ( see section [ opacity ] ) and because these lines may also be tracing more external layers of the clumps . in the case of sio , and also hco@xmath1 , the scatter is probably produced by molecular outflow activity , which increases the @xmath41v of the lines and adds extended wing emission . c@xmath0h appears to present the best correlation with n@xmath0h@xmath1 . h@xmath2co@xmath1 , hn@xmath2c , and hc@xmath3n have slightly narrower line widths than n@xmath0h@xmath1 , indicating that they trace similar optically thin gas emanating from the center of the clumps . from figure [ fwhm - plot ] , it can be seen that there are four sources with very large linewidths in the hco@xmath1 ( @xmath438.0 ) and sio ( @xmath436.0 ) lines . they are g018.82 mm3 , g018.82 mm4 , g035.59 mm1 , and g033.69 mm2 for hco@xmath1 and g018.82 mm3 , g019.27 mm2 , g022.35 mm1 , and g028.04 mm1 for sio . notably , the four sources that have large sio linewidths are active clumps . @xcite mention that larger sio linewidths and abundances are expected in younger sources than msx sources ( our red clumps ) . however , as they also mention , strong shocks are necessary to produce sio in the gas phase ; therefore , some star formation activity is needed and the sources should be older than quiescent clumps . because the n@xmath0h@xmath1 @xmath29 line has a detection rate of 97% and shows no self - absorbed profiles , we used the n@xmath0h@xmath1 systemic velocities of the 155 detected clumps from the initial sample to estimate kinematic distances . the distances were obtained using the @xcite rotation curve of the milky way , scaled to @xmath44 and included a small velocity correction ( 7 ) accounting for the measured solar peculiar motion . the distances of all clumps are listed in table [ tbl - coor ] . to resolve the ambiguity between the near and far kinematic distances , we used the fact that irdcs are seen as dark extinction features against the bright mid infrared background emission . hence , the ambiguity is resolved by assuming that all irdcs are located at the near kinematic distance ( e.g. , * ? ? ? * ; * ? ? ? the distances to the irdc clumps in this work were previously established by @xcite . they carried out a morphological matching of the _ msx _ mid infrared extinction with the @xmath2co @xmath31 molecular line emission from the boston university - five college radio astronomy observatory ( bu - fcrao ) galactic ring survey ( grs ; * ? ? ? however , because the @xmath2co @xmath31 emission is a low density tracer ( @xmath4510@xmath13 @xmath10 ) , there are typically several velocity components along one line of sight . of the initial 155 detected irdc clumps , the kinematic distance indicated by the high density tracer n@xmath0h@xmath1 ( @xmath4610@xmath9 @xmath10 ) of 35 ( 23% ) clumps differs from the distance inferred from @xmath2co by @xcite . of these sources , 17 show strong signatures of star formation and are classified as red and active clumps . because they are isolated clumps located apart from the irdc and at a different distance , they are not physically associated with the irdc and were excluded from further analysis in this paper . assuming local thermodynamic equilibrium ( lte ) conditions , the total column density , @xmath47 , can be derived from ( e.g. , * ? ? ? * ) @xmath48}\int \tau \,dv~ , \label{eqn - den - colum-2}\ ] ] where @xmath49 is the optical depth of the line , @xmath50 is the statistical weight of the upper level , @xmath51 is the einstein coefficient for spontaneous emission , @xmath52 is the energy of the lower state , @xmath53 is the partition function , @xmath54 is the transition frequency , and @xmath55 is the relative intensity of the brightest hyperfine transition with respect to the others . @xmath55 is only relevant for hyperfine transitions because it takes into account the satellite lines correcting by their relative opacities . it is 5/9 for n@xmath0h@xmath1 and 5/12 for c@xmath0h . it is equal to 1.0 for transitions without hyperfine structure . a particular case of equation [ eqn - den - colum-2 ] is the total column density of a linear , rigid rotor molecule , which is given by ( e.g. , * ? ? ? * ) @xmath56}\int\tau_{\nu}\,dv~ , \label{eqn - den - colum}\ ] ] where @xmath16 is the permanent dipole moment of the molecule , @xmath57 is the rotational quantum number of the lower state , @xmath58 is the energy in the level @xmath57 , and @xmath59 is the rotational constant of the molecule . to calculate column densities it is necessary to estimate the optical depths of the lines . we calculate the optical depths of n@xmath0h@xmath1 and c@xmath0h using their hyperfine structure , and the optical depths of hco@xmath1 and hnc using their isotopologues , h@xmath2co@xmath1 and hn@xmath2c . under the assumption of lte , all levels are populated according to the same excitation temperature ( @xmath60 ) . we assume that @xmath60 is equal to the dust temperature ( @xmath61 ) obtained by @xcite . they determined @xmath61 for 59 sources and gave lower and upper limits for 23 other sources in our sample of 92 irdc clumps by fitting a graybody function to the spectral energy distribution ( sed ) . for the sources with lower and upper limits , the average value of the limits was used as @xmath61 . for the remaining 10 sources , without seds , we use the median values found by @xcite for each evolutionary sequence : 23.7 , 30.4 , 34.5 , and 40.4 k for quiescent , intermediate , active , and red clumps , respectively . when the weakest line needed to compute the optical depth was not detected above the 3@xmath25 level ( n@xmath0h@xmath1 @xmath62 , c@xmath0h @xmath63 , h@xmath2co@xmath1 , or hn@xmath2c ) , the intensity of the line was assumed to be 3@xmath26 . this gives an upper limit of the line intensity and thus an upper limit in the optical depth . actual optical depths were determined as follows . the main beam brightness temperature ( @xmath64 ) is the antenna temperature ( @xmath65 ) corrected by the main bean efficiency ( @xmath66 ) , @xmath67 . the main beam brightness temperature of a line is related to the excitation temperature and the optical depth by @xmath68(1 - e^{-\tau_{\nu}})~ , \label{tmb - eq}\ ] ] where @xmath69 is the filling factor , @xmath49 is the optical depth of the line , @xmath70 is the background temperature , and @xmath71 is defined by @xmath72 taking the ratio of equation [ tmb - eq ] evaluated for two hyperfine components or two isotopologues and assuming that the filling factors are the same , we obtain @xmath73~ , \label{tau - eqn2}\ ] ] where the subscripts `` @xmath74 '' and `` @xmath75 '' refer to two different hyperfine components or isotopologue molecules . defining `` @xmath76 '' as the ratio between the optical depths , @xmath77 , and utilizing the fact that the term between brackets is @xmath61 , equation [ tau - eqn2 ] becomes @xmath78 solving this equation numerically , we obtain the optical depth for one of the components , and by using the relationship between opacities we can obtain the optical depths of the second component . the opacities of the n@xmath0h@xmath1 hyperfine components are obtained from the ratio between the observed main beam brightness temperatures of the two brightest transitions : @xmath29 and @xmath62 . rewriting equation [ tau - eqn ] , we obtain @xmath79 where the subscripts `` @xmath80 '' and `` @xmath81 '' refer to the n@xmath0h@xmath1 hyperfine components @xmath29 and @xmath62 , respectively . we have also used the opacity ratio between these two hyperfine lines , @xmath76=@xmath82=@xmath83 , which depends only on the transition moments . the n@xmath0h@xmath1 opacities were calculated for 63 irdc clumps and the values for the main component are listed in table [ tbl - tau - column ] . the main component of n@xmath0h@xmath1 has opacities less than 1.0 in roughly 60% of the sample . the other 40% of the sample is moderately optically thick , with a maximum opacity of 4.6 . the median values for quiescent , intermediate , active , and red clumps are 0.8 , 0.8 , 0.5 , and 1.8 , respectively . the optical depths of the c@xmath0h hyperfine components are obtained using the intensity ratio between the @xmath30 and @xmath63 transitions . the opacity ratio between these two lines is @xmath76=@xmath84=2 @xcite . thus , the opacity of the brightest component is given by @xmath85 where the subscripts `` @xmath86 '' and `` @xmath87 '' refer to the c@xmath0h hyperfine components @xmath30 and @xmath63 , respectively . the c@xmath0h opacities were calculated for 43 irdc clumps and the values for the brightest transition are listed in table [ tbl - tau - column ] . the emission of the brightest c@xmath0h hyperfine component is mostly moderately optically thick , with 86% of the optical depths ranging from 1.0 to 6.0 . the median values of the optical depths for quiescent , intermediate , active , and red clumps are 2.9 , 2.4 , 2.2 , and 1.4 , respectively . rewriting equation [ tau - eqn ] , we obtain @xmath88 where `` @xmath76 '' is the ratio between the optical depths , @xmath89 , and the subscripts `` @xmath90 '' and `` @xmath91 '' refer to the hco@xmath1 and h@xmath2co@xmath1 isotopologues , respectively . a way to determine `` @xmath76 '' is to take the ratio of the expression for the column density ( see equation [ eqn - den - colum ] ) evaluated for both isotopologues @xmath92\frac{(kt_{\rm ex}/hb_{13 } + 1/3)}{(kt_{\rm ex}/hb_{12 } + 1/3)}\frac{\exp ( e_{_{j_{13}}}/kt_{\rm ex})}{\exp(e_{_{j_{12}}}/kt_{\rm ex})}\times \nonumber\\ & & \frac{[1 - \exp ( -h\nu_{_{12}}/kt_{\rm ex})]}{[1 - \exp(-h\nu_{_{13}}/kt_{\rm ex})]}~ , \label{r - hco+_tau}\end{aligned}\ ] ] where [ hco@xmath1/h@xmath2co@xmath1 ] is the isotopic abundance ratio . the isotopic abundance ratio of [ @xmath94c/@xmath2c ] ranges from @xmath620 to @xmath670 ( e.g. , * ? ? ? assuming a constant [ hco@xmath1/h@xmath2co@xmath1 ] abundance ratio of 50 for all sources , we compute the optical depths of 62 irdc clumps . the values for hco@xmath1 are summarized in table [ tbl - tau - column ] . all sources have optically thick hco@xmath1 emission , with opacities ranging between 7 and 75 . on the other hand , h@xmath2co@xmath1 is optically thin in most of the sources ( 90% ) . the median values of hco@xmath1 optical depths for quiescent , intermediate , active , and red clumps are 30 , 20 , 16 , and 19 , respectively ; and the values for h@xmath2co@xmath1 are 0.6 , 0.4 , 0.3 , and 0.4 , respectively . rewriting equation [ tau - eqn ] and following the same procedure used for hco@xmath1 , we obtain @xmath95 where @xmath76 is the ratio between the optical depths , @xmath89 , and is given by @xmath96\frac{(kt_{\rm ex}/hb_{13 } + 1/3)}{(kt_{\rm ex}/hb_{12 } + 1/3)}\frac{\exp ( e_{_{j_{13}}}/kt_{\rm ex})}{\exp(e_{_{j_{12}}}/kt_{\rm ex})}\times \nonumber\\ & & \frac{[1 - \exp ( -h\nu_{_{12}}/kt_{\rm ex})]}{[1 - \exp(-h\nu_{_{13}}/kt_{\rm ex})]}~ , \label{r - hco+_tau}\end{aligned}\ ] ] where the subscripts `` @xmath90 '' and `` @xmath91 '' refer to the hnc and hn@xmath2c isotopologues , respectively , and [ hnc / hn@xmath2c ] is the isotopic abundance ratio . assuming an [ hnc / hn@xmath2c ] abundance ratio of 50 , we compute the optical depths of 85 irdc clumps . the values for hnc are summarized in table [ tbl - tau - column ] . all sources have optically thick hnc emission , whereas hn@xmath2c is optically thin in most of the sources ( 95% ) . the median values of hnc optical depths for quiescent , intermediate , active , and red clumps are 24 , 22 , 18 , and 16 , respectively ; and for hn@xmath2c are 0.5 , 0.4 , 0.3 , and 0.3 , respectively . lccccccccccc g015.05 mm1 & 1.00 & 29.9 & 26.9 & 2.8 & 1.56(0.12 ) & 2.05(0.72 ) & 2.27(1.09 ) & 2.75(3.02 ) & ... & 1.67(0.69 ) & ... + g015.05 mm2 & 0.04 & ... & 59.4 & ... & 0.71(0.15 ) & 0.58(0.21 ) & 2.66(1.91 ) & ... & ... & ... & ... + g015.05 mm3 & 0.60 & ... & 28.6 & ... & 0.36(0.10 ) & ... & 1.04(0.51 ) & ... & ... & ... & ... + g015.05 mm4 & ... & ... & 16.1 & 4.5 & 1.04(0.12 ) & 0.58(0.20 ) & 1.56(0.65 ) & 3.06(3.29 ) & ... & ... & ... + g015.05 mm5 & 0.36 & 35.0 & 52.2 & ... & 0.52(0.16 ) & 1.01(0.56 ) & 2.40(1.55 ) & ... & ... & ... & ... + g015.31 mm2 & 2.89 & ... & 69.0 & ... & 1.18(1.25 ) & ... & 2.19(1.79 ) & ... & ... & ... & ... + g015.31 mm3 & ... & 50.2 & 30.3 & ... & 0.31(0.08 ) & 0.85(0.61 ) & 0.51(0.51 ) & ... & ... & ... & ... + g015.31 mm5 & ... & 38.2 & 53.0 & ... & ... & 1.10(0.63 ) & 1.36(0.95 ) & ... & ... & ... & ... + g018.82 mm2 & 2.16 & 44.6 & 43.5 & ... & 2.24(2.25 ) & 3.20(1.94 ) & 3.45(2.10 ) & ... & ... & ... & ... + g018.82 mm3 & 2.27 & 37.7 & 27.9 & ... & 6.07(6.15 ) & 10.10(5.24 ) & 6.71(3.14 ) & ... & ... & ... & ... + g018.82 mm4 & 0.12 & 75.0 & 45.3 & ... & 1.06(0.18 ) & 8.50(6.72 ) & 5.02(2.85 ) & ... & ... & 0.98(1.04 ) & ... + g018.82 mm6 & 0.12 & ... & ... & ... & 1.22(0.12 ) & 0.52(0.19 ) & 0.80(0.35 ) & ... & ... & ... & ... + g019.27 mm2 & 0.81 & ... & 33.8 & 5.6 & 2.18(0.15 ) & ... & 4.09(1.40 ) & 4.15(4.59 ) & ... & ... & ... + g022.35 mm1 & 0.31 & 14.4 & 20.9 & ... & 0.40(0.10 ) & 0.85(0.33 ) & 1.13(0.47 ) & ... & ... & ... & ... + g022.35 mm2 & ... & ... & 33.2 & ... & ... & ... & 2.02(1.07 ) & ... & ... & ... & ... + g023.60 mm7 & 1.64 & ... & 27.6 & ... & 4.13(4.07 ) & ... & 3.50(1.67 ) & ... & ... & ... & ... + g023.60 mm9 & 0.56 & ... & 29.3 & 2.4 & 0.68(0.11 ) & ... & 1.47(0.71 ) & 1.51(2.16 ) & ... & ... & ... + g024.08 mm2 & 0.99 & 43.0 & 26.4 & ... & 0.70(0.12 ) & 1.77(0.71 ) & 2.51(1.12 ) & ... & ... & ... & ... + g024.08 mm3 & 1.27 & 31.5 & 24.1 & ... & 0.93(0.92 ) & 1.78(0.89 ) & 1.91(0.85 ) & ... & ... & ... & ... + g024.08 mm4 & 0.27 & 66.5 & 31.1 & ... & 0.48(0.13 ) & 2.43(2.20 ) & 1.84(0.94 ) & ... & ... & ... & ... + g024.33 mm2 & ... & ... & 25.5 & ... & 3.06(0.18 ) & 1.37(0.36 ) & 6.16(1.17 ) & ... & ... & 4.27(1.10 ) & ... + g024.33 mm3 & ... & ... & ... & 1.9 & 2.65(0.17 ) & 1.38(0.35 ) & 0.85(0.29 ) & 3.41(4.60 ) & ... & 2.81(0.90 ) & 3.96(1.41 ) + g024.33 mm4 & ... & ... & ... & ... & 1.57(0.13 ) & 0.79(0.30 ) & 0.78(0.28 ) & ... & ... & ... & ... + g024.33 mm5 & 0.26 & ... & ... & 3.2 & 2.17(0.15 ) & 2.81(0.46 ) & 0.79(0.29 ) & 2.63(3.99 ) & ... & 5.44(0.94 ) & ... + g024.33 mm7 & 0.67 & 15.0 & 10.3 & 1.2 & 0.93(0.09 ) & 1.22(0.23 ) & 1.06(0.39 ) & 0.61(1.20 ) & ... & ... & 1.36(0.74 ) + g024.33 mm8 & ... & 13.2 & 15.8 & ... & ... & 1.40(0.53 ) & 2.62(1.01 ) & ... & ... & ... & ... + g024.33 mm9 & ... & 10.9 & 11.5 & ... & ... & 2.30(0.89 ) & 3.48(1.36 ) & ... & ... & ... & ... + g024.33 mm11 & 0.59 & 29.9 & 23.9 & ... & 1.51(0.10 ) & 2.82(0.62 ) & 4.01(1.49 ) & ... & ... & ... & ... + g024.60 mm2 & 0.92 & 22.4 & 16.4 & 1.1 & 1.24(0.10 ) & 1.51(0.49 ) & 1.48(0.63 ) & 1.16(0.40 ) & 0.35(0.12 ) & ... & ... + g025.04 mm2 & ... & 13.2 & 18.3 & ... & 0.63(0.11 ) & 1.26(0.51 ) & 1.70(0.73 ) & ... & ... & ... & ... + g025.04 mm4 & 0.06 & 8.7 & 24.6 & 1.0 & 2.44(0.11 ) & 1.51(0.54 ) & 4.30(1.01 ) & 1.09(2.22 ) & ... & 3.12(0.66 ) & ... + g027.75 mm2 & 3.17 & ... & ... & ... & 1.01(1.02 ) & ... & ... & ... & ... & ... & ... + g027.94 mm1 & 1.39 & ... & 22.9 & 3.1 & 1.66(1.66 ) & ... & 1.97(0.68 ) & 1.94(3.13 ) & ... & ... & ... + g028.04 mm1 & 1.15 & 24.1 & 60.7 & 1.3 & 2.27(2.23 ) & 2.24(0.96 ) & 5.50(2.66 ) & 1.27(2.27 ) & 0.38(0.12 ) & 2.88(0.84 ) & ... + g028.08 mm1 & 1.25 & ... & 40.7 & ... & 0.63(0.63 ) & ... & 1.92(1.02 ) & ... & ... & ... & ... + g028.23 mm1 & 3.62 & ... & ... & ... & 3.12(3.14 ) & 1.00(0.34 ) & 0.77(0.27 ) & ... & ... & ... & ... + g028.28 mm4 & 0.88 & 13.9 & 11.1 & 3.0 & 1.53(0.13 ) & 1.83(0.38 ) & 1.65(0.60 ) & 2.68(3.09 ) & ... & ... & ... + g028.37 mm1 & ... & ... & 20.6 & 1.9 & 4.60(0.15 ) & 2.51(0.39 ) & 5.72(0.97 ) & 4.31(2.38 ) & 1.45(0.14 ) & 9.93(0.88 ) & 16.00(1.36 ) + g028.37 mm2 & 0.46 & ... & 24.6 & 3.0 & 2.51(0.18 ) & 2.42(0.45 ) & 6.37(1.76 ) & 4.54(6.44 ) & 0.49(0.21 ) & 3.61(1.50 ) & ... + g028.37 mm4 & 0.06 & 12.8 & 11.8 & 4.8 & 4.69(0.19 ) & 2.90(0.30 ) & 4.37(0.47 ) & 6.30(5.60 ) & 0.85(0.11 ) & 8.05(0.63 ) & 13.90(0.98 ) + g028.37 mm6 & 0.19 & 22.2 & 20.3 & 2.3 & 2.67(0.10 ) & 2.72(0.52 ) & 2.88(0.46 ) & 1.80(1.89 ) & 0.77(0.12 ) & 4.30(0.51 ) & 10.20(0.90 ) + g028.37 mm9 & ... & 21.9 & 15.0 & ... & 1.52(0.11 ) & 1.99(0.34 ) & 1.52(0.29 ) & ... & ... & ... & ... + g028.37 mm11 & 0.23 & ... & 41.8 & 3.3 & 1.05(0.20 ) & 0.82(0.28 ) & 7.21(3.80 ) & 2.63(4.22 ) & ... & ... & ... + g028.37 mm12 & 1.37 & 65.8 & 74.0 & ... & 0.89(0.90 ) & 2.42(1.38 ) & 3.61(2.86 ) & ... & 0.31(0.15 ) & ... & ... + g028.37 mm13 & 3.32 & ... & ... & ... & 1.46(1.58 ) & ... & ... & ... & ... & ... & ... + g028.53 mm3 & 1.73 & 19.4 & 23.7 & ... & 2.47(2.42 ) & 2.13(0.64 ) & 2.38(1.04 ) & ... & ... & ... & ... + g028.53 mm5 & 4.45 & 46.0 & 46.7 & ... & 2.90(3.03 ) & 1.50(1.02 ) & 3.43(2.01 ) & ... & ... & ... & ... + g028.53 mm7 & ... & ... & 49.0 & ... & 1.10(0.12 ) & 0.91(0.26 ) & 3.47(1.21 ) & ... & ... & ... & ... + g028.53 mm8 & ... & 69.4 & 37.1 & 1.5 & 1.00(0.14 ) & 3.54(2.78 ) & 3.44(1.77 ) & 2.02(1.15 ) & ... & ... & ... + g028.53 mm9 & 0.81 & 13.9 & 15.0 & ... & 1.88(0.16 ) & 2.38(0.95 ) & 3.08(1.01 ) & ... & ... & ... & ... + g028.53 mm10 & 0.10 & 13.7 & 17.2 & 2.9 & 2.04(0.20 ) & 2.83(1.08 ) & 4.90(1.46 ) & 5.05(6.87 ) & ... & ... & ... + g028.67 mm1 & ... & 31.0 & 31.2 & ... & ... & 1.64(0.82 ) & 2.95(1.41 ) & ... & ... & ... & ... + g028.67 mm2 & 0.83 & ... & 47.7 & ... & 0.71(0.10 ) & ... & 2.64(1.57 ) & ... & ... & ... & ... + g030.14 mm1 & 4.58 & 37.8 & 32.6 & ... & 1.36(1.50 ) & 1.18(0.68 ) & 1.61(0.83 ) & ... & ... & ... & ... + g030.57 mm1 & 0.11 & ... & 19.6 & 1.1 & 4.42(0.15 ) & ... & 3.85(1.22 ) & 2.64(4.13 ) & 0.85(0.15 ) & 2.86(0.98 ) & ... + g030.57 mm3 & 0.42 & ... & 45.2 & 6.0 & 0.98(0.18 ) & ... & 6.33(3.59 ) & 2.39(3.67 ) & ... & ... & ... + g030.97 mm1 & ... & 16.4 & 9.4 & ... & 3.64(0.14 ) & 5.17(0.50 ) & 4.04(0.67 ) & ... & 0.37(0.12 ) & ... & ... + g031.97 mm5 & 0.04 & ... & 15.2 & 3.5 & 2.33(0.13 ) & 1.25(0.29 ) & 3.53(1.15 ) & 2.93(3.80 ) & ... & ... & ... + g031.97 mm7 & 0.45 & 7.3 & 14.5 & 3.5 & 1.69(0.10 ) & 1.20(0.23 ) & 3.00(2.81 ) & 2.14(3.40 ) & ... & 1.00(0.56 ) & ... + g031.97 mm8 & ... & ... & 9.5 & ... & 3.43(0.20 ) & ... & 3.41(1.18 ) & ... & ... & ... & ... + g033.69 mm1 & ... & 9.8 & 11.4 & ... & 3.56(0.11 ) & 1.70(0.25 ) & 2.93(0.34 ) & ... & 0.96(0.13 ) & 4.55(0.59 ) & 5.29(1.01 ) + g033.69 mm2 & ... & 17.1 & 14.6 & 0.8 & 3.19(0.21 ) & 4.53(1.84 ) & 4.48(1.14 ) & 3.55(0.49 ) & 1.01(0.16 ) & 4.31(1.21 ) & ... + g033.69 mm3 & ... & 18.5 & 16.4 & ... & 2.40(0.28 ) & 3.61(1.51 ) & 5.18(2.09 ) & ... & ... & ... & ... + g033.69 mm4 & ... & 12.4 & 16.2 & ... & 3.05(0.21 ) & 1.76(0.64 ) & 3.37(0.83 ) & ... & 0.56(0.20 ) & 3.07(1.02 ) & ... + g033.69 mm5 & ... & 31.6 & 18.3 & 2.9 & 2.15(0.24 ) & 3.85(1.91 ) & 4.23(1.76 ) & 3.75(3.05 ) & ... & ... & ... + g033.69 mm11 & ... & 12.1 & 13.9 & 4.0 & ... & 2.53(0.95 ) & 3.99(1.52 ) & 1.48(2.87 ) & ... & ... & ... + g034.43 mm1 & 0.42 & ... & 18.0 & 1.7 & 13.70(0.21 ) & 4.77(0.48 ) & 13.60(1.58 ) & 7.75(4.71 ) & 3.33(0.21 ) & 8.59(1.43 ) & 34.70(2.09 ) + g034.43 mm5 & 0.67 & ... & 30.0 & ... & 2.98(0.09 ) & 1.92(0.25 ) & 5.23(0.79 ) & ... & 0.54(0.12 ) & ... & ... + g034.43 mm7 & ... & 21.6 & 18.9 & 1.9 & 1.11(0.11 ) & 1.07(0.33 ) & 2.29(0.60 ) & 2.06(2.92 ) & 0.21(0.12 ) & ... & ... + g034.43 mm8 & ... & 65.7 & 19.0 & 3.9 & 1.53(0.17 ) & 2.71(1.21 ) & 2.55(1.01 ) & 3.56(5.98 ) & ... & ... & ... + g034.77 mm1 & 2.20 & 27.9 & 11.2 & ... & 4.09(4.04 ) & 4.52(0.79 ) & 2.44(0.80 ) & ... & 0.47(0.13 ) & ... & ... + g034.77 mm3 & 0.25 & ... & 22.4 & 4.1 & 0.88(0.15 ) & 0.50(0.18 ) & 2.78(1.17 ) & 3.01(5.16 ) & ... & ... & ... + g035.39 mm7 & 0.18 & 27.9 & 17.6 & 2.1 & 2.41(0.17 ) & 6.00(0.94 ) & 3.81(0.82 ) & 4.38(4.13 ) & 0.32(0.15 ) & ... & ... + g035.59 mm1 & 4.18 & 31.0 & 40.8 & ... & 2.84(3.10 ) & 4.38(2.19 ) & 1.77(1.10 ) & ... & ... & ... & ... + g035.59 mm2 & ... & 21.8 & 11.6 & ... & 0.72(0.10 ) & 1.82(0.57 ) & 1.09(0.41 ) & ... & ... & ... & ... + g035.59 mm3 & 0.42 & 7.8 & 10.5 & 2.4 & 1.00(0.23 ) & 1.12(0.29 ) & 1.25(0.47 ) & 1.82(2.20 ) & 0.14(0.19 ) & ... & ... + g036.67 mm1 & 3.22 & 68.4 & 19.7 & 4.3 & 0.99(1.02 ) & 1.58(1.34 ) & 1.02(0.60 ) & 1.58(1.57 ) & ... & ... & ... + g036.67 mm2 & ... & 20.2 & 18.1 & ... & ... & 0.65(0.28 ) & 1.15(0.48 ) & ... & ... & ... & ... + g038.95 mm1 & 0.12 & 26.9 & 9.6 & 1.2 & 1.34(0.10 ) & 4.27(0.65 ) & 1.69(0.51 ) & 2.41(1.94 ) & 0.47(0.19 ) & 0.92(0.53 ) & ... + g038.95 mm2 & 0.92 & 19.7 & 11.1 & ... & 1.72(0.20 ) & 4.64(1.06 ) & 2.45(0.92 ) & ... & 0.35(0.17 ) & ... & ... + g038.95 mm3 & 2.44 & 18.5 & 14.0 & 0.4 & 2.98(2.97 ) & 3.40(0.87 ) & 2.35(0.92 ) & 2.80(0.40 ) & ... & ... & ... + g038.95 mm4 & 1.66 & 10.4 & 13.8 & 1.0 & 1.97(1.99 ) & 1.83(0.68 ) & 2.34(0.92 ) & 2.16(0.39 ) & ... & ... & 2.48(1.80 ) + g053.11 mm1 & 0.56 & 10.5 & 4.5 & ... & 3.11(0.25 ) & 4.74(0.48 ) & 1.97(0.59 ) & ... & 0.80(0.14 ) & ... & ... + g053.11 mm2 & 1.05 & 13.6 & 9.6 & 1.1 & 2.76(2.77 ) & 2.80(0.56 ) & 1.87(0.53 ) & 1.62(0.62 ) & ... & ... & ... + g053.11 mm4 & 0.52 & 23.0 & 7.1 & 0.2 & 1.11(0.17 ) & 4.03(0.61 ) & 1.45(0.39 ) & 0.94(0.36 ) & 0.16(0.12 ) & ... & ... + g053.11 mm5 & 4.24 & 11.0 & 12.4 & 0.9 & 2.90(2.96 ) & 1.26(0.47 ) & 1.54(0.58 ) & 1.08(0.25 ) & ... & ... & ... + g053.25 mm1 & ... & 7.8 & 5.7 & 1.5 & 1.44(0.14 ) & 1.38(0.45 ) & 1.20(0.42 ) & 1.54(2.06 ) & ... & ... & ... + g053.25 mm3 & 1.79 & 15.3 & 10.4 & 4.4 & 1.37(1.44 ) & 1.49(0.62 ) & 1.35(0.52 ) & 3.10(5.18 ) & ... & ... & ... + g053.25 mm4 & 0.25 & 12.4 & 5.8 & ... & 1.62(0.15 ) & 2.44(0.56 ) & 1.23(0.43 ) & ... & ... & ... & ... + g053.25 mm5 & ... & 21.4 & 10.7 & 4.8 & ... & 2.55(0.88 ) & 1.38(0.52 ) & 2.05(4.63 ) & ... & ... & ... + g053.25 mm6 & 1.46 & 13.3 & 7.5 & ... & 3.84(3.88 ) & 2.64(0.62 ) & 1.65(0.62 ) & ... & ... & ... & ... + g053.31 mm2 & 2.69 & 11.2 & 13.3 & ... & 1.24(1.23 ) & 0.67(0.25 ) & 1.00(0.39 ) & ... & ... & ... & ... + the column densities of n@xmath0h@xmath1 , hco@xmath1 and hnc are calculated using equation [ eqn - den - colum ] . the sio molecule is also a linear , rigid rotor . thus , we can also use equation [ eqn - den - colum ] to estimate sio column densities . however , since we observed only one sio isotopologue , we need to assume that the sio emission is optically thin and the filling factor is 1.0 ( which gives us the beam - averaged column density ) . therefore , replacing equation [ tmb - eq ] in equation [ eqn - den - colum ] , we obtain @xmath97}\times \nonumber\\ & & \frac{1}{[j(t_{\rm ex } ) - j(t_{\rm bg})]}\int t_{\rm mb}\,dv~. \label{eqn - den - colum - thin}\end{aligned}\ ] ] the values of the parameters used in the column density calculations ( permanent dipole moment , rotational constant ) of n@xmath0h@xmath1 , hco@xmath1 , hnc , and sio are summarized in table [ tbl - param - column1 ] . lccl & 3.40 & 46.586871 & 1 , 2 , 3 + & 3.89 & 44.594423 & 4 , 5 , 6 + & 3.89 & 43.377302 & 4 , 5 , 6 + & 3.05 & 45.331980 & 7 , 8 + & 3.05 & 43.545600 & 7 , 8 + & 3.10 & 21.711979 & 9 , 10 + the column densities of the remaining molecules ( c@xmath0h , hc@xmath3n , hnco ) are determined using equation [ eqn - den - colum-2 ] . although c@xmath0h is a linear molecule , its rotational energy levels are described by the rotational quantum number @xmath47 , instead of @xmath57 . hc@xmath3n is also linear and its partition function is the same used for diatomic , linear molecules @xcite ; we have not included the factor 3 due to the nuclear spin degeneracy because the hyperfine structure is not resolved . on the other hand , hnco is a standard asymmetric top molecule with a different partition function @xcite . for c@xmath0h the column density was calculated using the optical depths obtained in section [ opacity ] . for hc@xmath3n and hnco , the emission was assumed to be optically thin and the filling factor was used to be equal to 1.0 . the values of the parameters used in the column density calculations ( partition function , rotational constants , statistical weight , einstein coefficient for spontaneous emission , and energy of the lower state ) of c@xmath0h , hc@xmath3n and hnco are summarized in table [ tbl - param - column2 ] . upper limits to the column densities of the most commonly detected molecular lines ( n@xmath0h@xmath1 , hco@xmath1 , hnc ) were estimated when they were not detected over the 3@xmath25 level . in the case of the optically thick hco@xmath98 and hnc emission lines , the limits were obtained first for their isotopologues ( h@xmath2co@xmath98 and hn@xmath2c ) . then by using the isotopic abundance ratio we calculated the column densities for hco@xmath98 and hnc . in all non - detections , the emission was assumed to be optically thin , with @xmath99 , and @xmath41v equal to 2.0 . additionally , lower limits to the column densities were estimated for n@xmath0h@xmath1 and hco@xmath1 when our normal approach was not applicable . for n@xmath0h@xmath1 , the intensity ratio between the two brightest components , which is used to derive opacities , should not be higher than its value in the optically thin limit , 5/3 . however , in some cases , this ratio exceeds the optically thin limit , but nevertheless remains consistent with 5/3 within the uncertainties . to calculate a lower limit for the column density , the main n@xmath0h@xmath1 component was assumed to be optically thin . for hco@xmath1 , in several cases the spectra show self - absorbed profiles , and no gaussian fit was performed . lower limits to the column densities were computed using the h@xmath2co@xmath98 emission line and assuming that its emission is optically thin . then , using the [ hco@xmath1/h@xmath2co@xmath1 ] isotopic abundance ratio , we obtain the hco@xmath1 column densities . derived column densities using the optical depths obtained in section [ opacity ] were multiplied by the filling factor to get the beam - averaged column density . the filling factor was calculated from equation [ tmb - eq ] . column densities for n@xmath0h@xmath1 , hco@xmath1 , hnc , sio , c@xmath0h , hc@xmath3n , and hnco molecules ( including all limits ) are listed in the table [ tbl - tau - column ] . median values of the column densities for each evolutionary sequence ( excluding limits when the lines are not detected above the 3@xmath25 level ) are shown in table [ tbl - medians ] . h@xmath1 column densities for each evolutionary sequence . the name of the evolutionary stage is given on the top left corner of each panel . the vertical dashed lines indicate the median values of the column densities for each distribution . the diagonal patterns indicate the portions of the histograms that correspond to upper and lower limits ( see bottom panel , right side ) . the median values and the k - s test suggest that column densities increase with the evolution of the clumps . median values of the column densities are given in table [ tbl - medians ] . ] figure [ n2h+_column_density ] shows the number distributions of the n@xmath0h@xmath1 column densities for each evolutionary sequence . the k s test gives a probability of 0.09% that quiescent and active clump distributions originate from the same parent population . figure [ n2h+_column_density ] shows a trend of increasing column densities with the evolution of the clump , which is confirmed by the median values ( see table [ tbl - medians ] ) and the k s test probability . this trend does not necessary imply a change in the chemistry . an increase of the n@xmath0h@xmath1 column density could simply indicate that the total h@xmath0 column density increases while maintaining constant chemical composition . in sections [ chemistry ] and [ molecules ] we discuss the molecular abundances , which more definitively show chemical variations . this argument applies to all the molecules discussed in this section . @xcite obtained n@xmath0h@xmath1 column densities over a sample of 11 _ spitzer _ and _ msx _ dark objects ( no 24 and 8 @xmath16 m emission , respectively ) that we can compare with our quiescent clumps . @xcite obtained n@xmath0h@xmath1 column densities of 34 clumps , where massive stars have already formed , that we can compare with our active clumps . table [ tbl - comparison ] summarizes the results of previous works , which are in good agreement with our results . column densities for each evolutionary sequence . the name of the evolutionary stage is given on the top left corner of each panel . the vertical dashed lines indicate the median values of the column densities for each distribution . the diagonal patterns indicate the portions of the histograms that correspond to upper and lower limits ( see bottom panel , right side ) . the median values and the k - s test suggest that column densities increase with the evolution of the clumps . median values of the column densities are given in table [ tbl - medians ] . ] figure [ hco+_column_density ] shows the number distributions of the hco@xmath1 column densities for each evolutionary stage . the k s test gives us a probability of 2% that quiescent and active distributions originate from the same parent population . figure [ hco+_column_density ] shows a trend of increasing column densities with the evolution of the clump , which is supported by the median values ( see table [ tbl - medians ] ) and the k s test probability . @xcite and @xcite obtained h@xmath2co@xmath1 and hco@xmath1 column densities toward 20 and 79 massive clumps , respectively , associated with ir emission , methanol masers and uc hii regions . we compare this sample with our active and red cores in table [ tbl - comparison ] . the high hco@xmath1 column densities obtained by @xcite and @xcite confirm the trend that hco@xmath1 column densities increase with the evolution of the clumps , as seen in figure [ hco+_column_density ] . figure [ hnc_column_density ] shows the number distributions of the hnc column densities for each evolutionary stage . the k s test gives a probability of 5% that quiescent and active distributions originate from the same parent population . figure [ hnc_column_density ] shows a weaker trend ( compared with n@xmath0h@xmath1 and hco@xmath1 ) of increasing column densities from quiescent to active clumps . there are only a few systematic studies of hnc toward a large sample of massive star forming or irdc clumps . @xcite observed hn@xmath2c toward 20 massive clumps that we can compare with our active clumps . their values , showed in table [ tbl - comparison ] , are in good agreement with our results . . ] figure [ c2h_column_density ] shows the number distributions of the c@xmath0h column densities for each evolutionary sequence . the k s test gives us a probability of 38% that active and quiescent distributions originate from the same parent population . figure [ c2h_column_density ] shows that the trend of increasing column densities with the evolution of the clumps is less evident for c@xmath0h than for previous molecules . although the median values obtained for each evolutionary sequence support this trend , the k s test probability is not low enough to be confident . table [ tbl - comparison ] compares the column densities obtained for 20 massive clumps by @xcite with our active clumps , which are in a reasonable agreement . the median values for the hc@xmath3n column densities are summarized in table [ tbl - medians ] . hc@xmath3n emission was detected only in one quiescent clump , g028.37 mm12 , whose column density was @xmath100 @xmath12 . @xcite obtained hc@xmath3n column densities for 7 _ spitzer _ and _ msx _ dark objects ( no 24 and 8 @xmath16 m emission , respectively ) . their values are comparable to that found in our one quiescent clump ( see table [ tbl - comparison ] ) . the median values for the hnco column densities are summarized in table [ tbl - medians ] . @xcite calculated hnco column densities toward 20 massive star forming clumps using rotational diagrams . these sources were selected based on the presence of water masers and strong sio emission , indicating they are currently forming stars . they obtained higher values than those found in our irdc sample ( table [ tbl - comparison ] ) . this may be explained by the warmer rotational temperatures ( median of 88 k ) obtained by @xcite . h column densities for each evolutionary sequence . the name of the evolutionary stage is given on the top left corner of each panel . the vertical dashed lines indicate the median values of the column densities for each distribution . the diagonal patterns indicate the portions of the histograms that correspond to upper limits . the median values slightly increase with the evolution of the clumps ; however , the k - s test shows this variation is not statistically significant . median values of the column densities are given in table [ tbl - medians ] . ] the median values for the sio column densities are summarized in table [ tbl - medians ] . sio emission was detected in no intermediate clumps and only in one quiescent clump , g024.33 mm7 , whose column density is @xmath101 @xmath12 . @xcite and @xcite obtained sio column densities for 17 and 15 high - mass star - forming clumps , respectively , associated with ir and maser emission and uc hii regions ( see table [ tbl - comparison ] ) . their values agree with those we calculated for active clumps . red clumps , however , have lower column densities . in order to study the chemistry toward the irdc clumps , and investigate if the different evolutionary sequences proposed by @xcite are chemically distinguishable , we calculate molecular abundances using the 1.2 mm continuum emission obtained by @xcite and abundance ratios of selected species . to estimate molecular abundances with respect to molecular hydrogen , we take the ratio between the column density of a given molecule and the h@xmath0 column density derived from dust emission . observations of 1.2 mm dust continuum emission were obtained by @xcite with an angular resolution of 11 . in order to calculate the molecular abundances , we smoothed the continuum emission data to the angular resolution of the molecular line data at 3 mm , 38 . the resulting 1.2 mm peak flux , h@xmath0 column densities and molecular abundances for the n@xmath0h@xmath1 , hco@xmath1 , hnc , c@xmath0h , hc@xmath3n , hnco , and sio lines ( including limits for n@xmath0h@xmath1 , hco@xmath1 , and hnc ) are listed in table [ tbl - mm - abun ] . in general , the number distributions of abundances show few differences between the evolutionary sequences , with the exception of n@xmath0h@xmath1 and hco@xmath1 ( figure [ n2h+_abundances ] and [ hco+_abundances ] ) . the k - s test gives a probability of 0.09% and 1.2% that quiescent and active distributions originate from the same parent populations for n@xmath0h@xmath1 and hco@xmath1 , respectively . on the other hand , the evolutionary status of the clumps is not evidently distinguishable from the abundances of hnc , c@xmath0h , hc@xmath3n , hnco , and sio , although their median values change with the evolution of the clumps . median values of the molecular abundances for each evolutionary sequence ( excluding the limits when the lines are not detected ) are shown in table [ tbl - medians ] . @xcite observed 37 irdc clumps in the fourth quadrant with the mopra telescope . they also obtained molecular abundances for n@xmath0h@xmath1 , hco@xmath1 , hnc , sio , c@xmath0h , hc@xmath3n , and hnco . despite a couple of different assumptions ( @xcite did not smooth their 1.2 mm dust continuum emission to the angular resolution of mopra ( 38 ) and they used nh@xmath3 tempearatures ) , the abundances they find are in good agreement with the abundances determined in this work , except for hnc . they determined hnc abundances about an order of magnitud lower than our values . because @xcite did not have the isotopologue hn@xmath2c , they assumed that the emission was optically thin in order to get an estimate of the column density . however , as we discussed in section [ opacity ] , the optically thin assumption is not valid for hnc in irdcs because the emission is optically thick . molecular abundance ratios that can be used to estimate the age and mark the evolutionary stages of star - forming regions are known as `` chemical clocks . '' only molecules that show differential abundances with time can be used to evaluate the evolutionary status of a star - forming region . chemical clocks have been studied in depth in low - mass star - forming regions ( e.g. , * ? ? ? * ) ; the concept of chemical clocks has been less developed in the context of high - mass star - forming regions . recent studies in small samples of high - mass star - forming regions show that chemical clocks can be extended to their high - mass counterpart ( e.g. , * ? ? ? * ) . in this section , we explore the idea that n@xmath0h@xmath1/hco@xmath1 and n@xmath0h@xmath1/hnc ratios can be used as chemical clocks . we choose this set of lines because they reveal the clearest significant chemical variations . lclccc c@xmath0h & @xmath102 & @xmath103 & 5.0 & 1.52757&0.00216 + hc@xmath3n&@xmath102 & @xmath104 & 21.0&58.1300&19.6484 + hnco & @xmath105^{1/2}$]&@xmath106&9.0 & 8.78011&6.32957 + & & @xmath107 & & & + & & @xmath108 & & & + it has been suggested that the line and abundance ratios of n@xmath0h@xmath1 and hco@xmath1 are anticorrelated , varying spatially over a given cloud and for different sources . the change in the n@xmath0h@xmath1/hco@xmath1 ratio was first noted by @xcite and @xcite , and recently by @xcite and @xcite ; however , it has not been studied in a large sample of star - forming clumps . although we can not examine spatial variations , due to our single - pointing observations , we can test whether the abundance ratios show differences for the four evolutionary stages proposed by @xcite . according to ion - molecule schemes of dense molecular clouds ( e.g. , * ? ? ? * ; * ? ? ? * ) , n@xmath0h@xmath1 and hco@xmath1 are formed primary through reactions with h@xmath109 : @xmath110 @xmath111 the main destruction mechanisms for n@xmath0h@xmath1 are electron recombination , when co is frozen out on dust grains , @xmath112 and by reacting with co , when co is found in the gas - phase at standard abundances ( @xmath113\sim 10^{-4}}$ ] ) , @xmath114 the dominant removal mechanism for hco@xmath1 is electron recombination @xmath115 chemical models of low - mass star - forming regions ( e.g. , * ? ? ? * ; * ? ? ? * ) and one of a massive star - forming region @xcite show that the n@xmath0h@xmath1/hco@xmath1 ratio is high in early stages due to hco@xmath1 depletion ( because most of the co is frozen out ) and is low in later stages when co is evaporated from the dust grains ( @xmath11620 - 25 k ) . because co is both the main supplier of hco@xmath1 and the main destroyer of n@xmath0h@xmath1 , the theory suggests that the hco@xmath1 abundance increases with respect to n@xmath0h@xmath1 as the clump evolves to a warmer phase . it is notable , for reasons not yet fully understood , that n@xmath0h@xmath1 is resistant to depletion in cold , dense regions . for a long time , it was thought that n@xmath0 , the main supplier of n@xmath0h@xmath1 , had a lower binding strength to the grain surface than co. however , recent measurements show that they are similar @xcite . hence , the chemical network that dominates the formation and destruction of n@xmath0h@xmath1 and hco@xmath1 in early stages of star formation is given by equations [ reac_n2h_form ] , [ reac_hco_form ] , [ reac_n2h_destruc1 ] and [ reac_hco_destruc ] , but for more evolved stages , equation [ reac_n2h_destruc1 ] becomes less important and is replaced by equation [ reac_n2h_destruc2 ] . a complete chemical network of h@xmath109 , n@xmath0h@xmath1 , and hco@xmath1 , including formation and destruction rates , is presented in detail by @xcite . figure [ n2h_hco_ratio ] shows the number distributions of the n@xmath0h@xmath1/hco@xmath1 abundance ( or column density ) ratio for each evolutionary sequence . the median values for each distribution are 0.08 , 0.12 , 0.10 , and 0.07 for quiescent , intermediate , active , and red clumps , respectively . as can be seen from the figure [ n2h_hco_ratio ] , quiescent clumps do not present the predicted abundance ratio to follow the trend of decreasing n@xmath0h@xmath1/hco@xmath1 ratio with the evolution of the clumps ; their ratios are more randomized . for this reason , we use the k - s test for the intermediate and red clump distributions . the test shows that the probability of both populations being the same is low ( 3% ) . the values of the median and the k - s test probability support the theoretical idea that the n@xmath0h@xmath1/hco@xmath1 abundance ratio acts as a chemical clock ( at least for the intermediate , active , and red clumps ) ; however , why this trend does not extend to quiescent clumps is unclear . we note that the relative abundance between n@xmath0h@xmath1 and hco@xmath1 acts as a chemical clock despite the abundances of both molecules independently increasing with the evolution of the clumps . we stress that a large number of values used in the histograms are limits and that a number of assumptions have gone into calculating the column densities and abundances . a sample with a larger number of strong detections would be very helpful for confirming these trends . the trend of decreasing n@xmath0h@xmath1/hco@xmath1 ratio with the evolution of the clumps may not extend to quiescent clumps due to one or a combination of the following reasons : ( a ) the sample of quiescent clumps could contain clumps with embedded star formation that was not detected by spitzer , ( b ) co has not had time to freeze out on dust grains , ( c ) the theoretical predictions at the very early stages are incorrect , and/or ( d ) due to the large beam of mopra , we are probably detecting contaminating , diffuse emission which is not directly related to the star - forming process in the center of the clumps . considering the last point , since the size of the mopra beam is 38 , we are typically observing emission on physical scales @xmath60.8 pc . typical sizes of clumps in high - mass star - forming regions and in irdcs are @xmath60.4 pc @xcite and @xmath60.5 pc @xcite , respectively . this shows that we are not only tracing the densest regions associated with star formation , but also the environment between clumps . @xcite predict that the variation of the molecular abundances is much larger when we consider small scales inside the clumps . further studies at higher angular resolution , using interferometers , are needed to understand the discrepancy shown by quiescent clumps . abundances for each evolutionary sequence . the name of the evolutionary stage is given on the top left corner of each panel . the vertical dashed lines indicate the median values of the abundances for each distribution . the diagonal patterns indicate the portions of the histograms that correspond to upper and lower limits ( see bottom panel , right side ) . the median values and the k - s test suggest that abundances increase with the evolution of the clumps , as it is also found by @xcite in their chemical modeling . median values of the column densities are given in table [ tbl - medians ] . ] figure [ n2h_hnc_ratio ] shows the n@xmath0h@xmath1/hnc abundance ratio for each evolutionary stage . the median values for each distribution are 0.06 , 0.06 , 0.08 , and 0.08 for quiescent , intermediate , active , and red clumps , respectively . the k - s test for the quiescent and active clump distributions shows that the probability of both populations being the same is low ( 5% ) . the median values and the k - s test suggest that there is a trend of increasing n@xmath0h@xmath1/hnc abundance ratio with the evolution of the clumps . we note that n@xmath0h@xmath1 and hco@xmath1 are strongly linked because of their pathways of formation and destruction ; on the other hand , a connection between n@xmath0h@xmath1 and hnc is less evident . the trend between these two molecules suggests that hnc may be preferentially formed in cold gas . it is expected that abundances of sio and more complex molecules , such as hc@xmath3n and hnco , increase in more evolved star - forming regions . in evolved regions , these molecules or their parent molecules are released from the dust grains by shocks and sublimation due to the temperature increase . in fact , the abundance ratio between each one of them and n@xmath0h@xmath1 , hco@xmath1 , or hnc show higher ratios in later stages of evolution . because the number of sio , hc@xmath3n and hnco detections is low , the histograms with the number distributions of abundance ratios were not made . on the other hand , c@xmath0h has a larger number of detections and shows no clear trends when abundance ratios are taken with respect to n@xmath0h@xmath1 , hco@xmath1 , or hnc . abundances for each evolutionary sequence . the name of the evolutionary stage is given on the top left corner of each panel . the vertical dashed lines indicate the median values of the abundances for each distribution . the diagonal patterns indicate the portions of the histograms that correspond to upper and lower limits ( see bottom panel , right side ) . the median values and the k - s test suggest that abundances increase with the evolution of the clumps . median values of the column densities are given in table [ tbl - medians ] . ] this section discusses the major assumptions and limitations that affect the calculations of column densities and molecular abundances . although the assumptions used in this work are widely used in the literature , it is important to be aware of their implications . * temperature . * under the assumption of local thermodynamic equilibrium ( lte ) conditions , we assume that the excitation temperature for all molecules is identical to the dust temperature for a given source . this assumption is reasonable since the critical densities of the molecules that were used to calculate physical parameters ( between @xmath117 - @xmath118 @xmath10 ) are comparable to the density of the sources ( @xmath610@xmath119 @xmath10 ; * ? ? ? * ) , implying they are close to thermalization . however , the dust temperature is an average temperature along the line of sight , and may not necessarily reflect the temperature of the dust and gas in the portion of the clumps / cores which are dense enough to produce the emission lines we are observing . we note that the molecular abundance ratios ( n@xmath0h@xmath1/hco@xmath1 and n@xmath0h@xmath1/hnc ) are practically independent of temperature because of a similar dependence on temperature for the individual abundances . the abudance ratio for a given source at 15 and 50 k varies only @xmath60.4% . * isotopic abundance ratio . * in the calculation of optical depths and column densities for hco@xmath1 and hnc , we have assumed a constant value of 50 for the [ @xmath94c/@xmath2c ] isotopic abundance ratio for all sources , regardless of their different distances to the galactic center . however , it has been shown that there is a gradient of this ratio increasing outwards from the galactic center @xcite . the [ @xmath94c/@xmath2c ] ratio has been measured using several methods and ranges from @xmath620 to @xmath670 ( * ? ? ? * and references therein ) . the variation of the [ @xmath94c/@xmath2c ] ratio can give an error of a factor @xmath62 in the column density estimation . * dust emission properties . * column densities were calculated using the smoothed 1.2 mm emission data from @xcite and the same procedure that they used . the main assumptions are that the dust opacity per gram of gas , @xmath120 , is 1.0 cm@xmath7 g@xmath122 , and the gas - to - dust mass ratio is 100 . the value for @xmath120 was determined by @xcite for protostellar cores assuming dust grains with thin ice mantles at gas densities of @xmath123 @xmath10 . however , dust opacities and the gas - to - dust ratio are not well constrained . they could vary due to dust destruction in later stages of evolution ( by shocks or uv radiation ) , or grain growth in dense , cold cores at very early stages . in the first scenario , the gas - to - dust ratio would be greater than 100 ; in the second scenario , the gas - to - dust ratio would be less than 100 and @xmath120 would decrease . h@xmath1/hco@xmath1 abundance ratios for each evolutionary sequence . the name of the evolutionary stage is given on the top left corner of each panel . the vertical dashed lines indicate the median values of the abundance ratios for each distribution . the diagonal patterns indicate the portions of the histograms that correspond to upper and lower limits ( see bottom panel , right side ) . the median values and the k - s test suggest that n@xmath0h@xmath1/hco@xmath1 ratios act as a chemical clock , increasing their values with the evolution of the clumps , from intermediate , to active , to red clumps . it is not clear why this trend does not include quiescent clumps ( see section [ abundance_ratios ] ) . median values of the abundance ratios are given in table [ tbl - medians ] . ] in this section , we discuss each commonly detected molecular line in depth . we describe the structure of the lines with hyperfine transitions ( n@xmath0h@xmath1 , hcn , and c@xmath0h ) , a brief review about their formation and previous works , and their behavior in irdc clumps . h@xmath1/hnc abundance ratios for each evolutionary sequence . the name of the evolutionary stage is given on the top left corner of each panel . the vertical dashed lines indicate the median values of the abundance ratios for each distribution . the diagonal patterns indicate the portions of the histograms that correspond to upper and lower limits ( see bottom panel , right side ) . the median values and the k - s test suggest that the n@xmath0h@xmath1/hnc ratio may act as chemical clock , increasing their median values with the evolution of the clumps , from quiescent to red clumps . median values of the abundance ratios are given in table [ tbl - medians ] . ] the n@xmath0h@xmath1 @xmath31 rotational transition exhibits a quadrupole hyperfine structure due to the spin of the two nitrogen nuclei . although for n@xmath0h@xmath1 @xmath31 there are 15 allowed hyperfine transitions @xcite , only seven features can be resolved in low - mass star forming regions ( e.g. , * ? ? ? * ; * ? ? ? * ) and three in high - mass star forming regions ( e.g. , * ? ? ? * ) . in the latter case , because massive star - forming regions show turbulent line widths ( @xmath433 ) much broader than thermal widths ( @xmath60.3 ) , two triplets are blended and observed as one line . the relative intensities in the optically thin case are given by the statistical weights 1:3:5:7 , when seven components are detected , and 1:3:5 , when three components are detected . the relative intensities become one in the optically thick case . n@xmath0h@xmath1 is a good tracer of dense , cold gas because it does not deplete at low temperatures and high densities . because the n@xmath0h@xmath1 spectrum show no evidence of self - absorbed profiles or line wings indicating outflow activity , this molecule is also a good tracer of dense , warm gas however , some clumps have broad lines , which may suggest they are formed by more than one central object . indeed , high angular resolution observations made with interferometers show the clumpy nature of irdcs ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? n@xmath0h@xmath1 is the second most frequently detected line in this survey and its detection rates do not depend of the star formation activity . however , the brightness temperature of n@xmath0h@xmath1 is normally weak in regions without signs of star formation . n@xmath0h@xmath1 column densities and abundances increase with the evolution of the clumps . although , in general , an increase of column density does not necessary mean an increase of the molecular abundance , we see in our sample of objects that column density and abundance of n@xmath0h@xmath1 tend to increase together . the increase of column density with the evolution of the clumps might be explained by the accretion of material . for instance , @xcite showed that , on average , clumps with signs of star formation have smaller sizes and higher densities than clumps without apparent star formation and @xcite propose that embedded protostars and protostellar cores undergo simultaneous mass growth during the protostellar evolution . the increase in the abundance might be explained by the rise in the temperature which releases n@xmath0 from the dust grains to the gas phase allowing the formation of more n@xmath0h@xmath1 with the passage of time . however , with the increase of temperature , co should also be released from the dust grains into the gas phase and should efficiently destroy n@xmath0h@xmath1 @xcite . why exactly n@xmath0h@xmath1 abundance increases with evolution is not clear . we suggest two possible explanations . first , the rates involving the formation and destruction of n@xmath0 , co , n@xmath0h@xmath1 , and hco@xmath1 might not be accurate , and n@xmath0h@xmath1 might not be destroyed as efficiently as is currently believed . second , the large beam of mopra is including emission from cold gas which is sorrounding the most compact regions directly associated with the star formation processes in the center of the clumps . if most of the n@xmath0h@xmath1 emission is coming from a cold envelope , in clumps that show signs of star formation , we are using in the abundance calculation a higher temperature than the actual value . higher temperatures would lead us to infer larger n@xmath0h@xmath1 abundances . @xcite performed a time - dependent chemical modeling of the massive protostellar cluster afgl5142 based on the model of @xcite . their model consists of a prestellar and a protostellar phase . from their figure 9 , it can be seen that once temperature increases , from the prestellar to the protostellar phase , the n@xmath0h@xmath1 molecular abundance also increases , supporting what we see in our work . hco@xmath1 has been widely used to investigate infall motions ( e.g. , * ? ? ? * ; * ? ? ? * ) , and , occasionally , its spectrum shows high - velocity wing emission which indicates outflow activity ( e.g. , * ? ? ? this makes hco@xmath1 a good tracer of kinematics in star - forming regions . on the other hand , h@xmath2co@xmath1 is optically thin and normally shows a gaussian profile . the combination of these two lines is usually used to distinguish between an asymmetric `` blue profile '' ( indicative of infall motions ) and two velocity components in the line of sight . if the optically thin h@xmath2co@xmath1 peaks in the self - absorption of the optically thick hco@xmath1 , it indicates a genuine `` blue profile . '' in our sample of irdc clumps , three intermediate clumps ( g024.33 mm2 , g024.33 mm5 , and g031.97 mm5 ) present this characteristic profile associated with large - scale infall motions . as also happpens for n@xmath0h@xmath1 , hco@xmath1 column densities and abundances rise with the evolution of the clumps . as we mentioned before for n@xmath0h@xmath1 , the increase in hco@xmath1 column densities might be due to the physical evolution of the clumps . the rise in hco@xmath1 abundance can be plausibly explained by the increase in the temperature , which releases co from grain mantles into the gas phase , forming more hco@xmath1 with the passage of time . for example , although @xcite focused on the nh@xmath3/n@xmath0h@xmath1 abundance ratio , they also obtained hco@xmath1 abundances from their chemical modeling . they showed that once temperature rises , from the prestellar to the protostellar phase , the hco@xmath1 molecular abundance rises , as well as that of n@xmath0h@xmath1 . lccccc + & 24 & 22 & 18 & 16 & 19 + & 0.8 & 0.8 & 0.5 & 1.8 & 0.8 + & 30 & 20 & 16 & 19 & 21 + & 2.9 & 2.4 & 2.2 & 1.4 & 2.4 + + & @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 & @xmath150 + & @xmath153 & ... & @xmath154 & @xmath155 & @xmath156 + + & @xmath157 & @xmath158 & @xmath159 & @xmath160 & @xmath161 + & @xmath162 & @xmath163 & @xmath164 & @xmath165 & @xmath166 + & @xmath167 & @xmath168 & @xmath169 & @xmath170 & @xmath171 + & @xmath172 & @xmath173 & @xmath174 & @xmath175 & @xmath176 + & @xmath177 & @xmath178 & @xmath179 & @xmath180 & @xmath180 + & @xmath181 & @xmath182 & @xmath183 & @xmath184 & @xmath185 + & @xmath186 & ... & @xmath187 & @xmath188 & @xmath189 + + n@xmath0h@xmath1/hco@xmath1 & 0.08 & 0.12 & 0.10 & 0.07 & 0.09 + n@xmath0h@xmath1/hnc & 0.06 & 0.06 & 0.08 & 0.08 & 0.07 + lccccccc this work & q : 1.03 & a : 2.51 & a : 3.39 & a : 2.66 & q : 3.05 & a : 4.30 & a : 1.39 + & a : 2.34 & r : 3.20 & ... & ... & ... & r : 4.31 & r : 0.46 + @xcite & q : 1.15 & ... & ... & ... & q : 9.70 & ... & ... + @xcite & a : 2.78 & ... & ... & ... & ... & ... & ... + @xcite & ... & a : 3.30 & a : 3.60 & a : 3.25 & ... & ... & a : 1.46 + @xcite & ... & r : 11.7 & ... & ... & ... & ... & ... + @xcite & ... & ... & ... & ... & ... & a - r : 8.61 & ... + @xcite & ... & ... & ... & ... & ... & ... & r : 2.00 + hnc is the molecule with the highest detection rate in the sample and its presence is independent of the star formation activity . because hnc is ubiquitous in the irdc clumps , and its spectrum shows no evidence of line wings indicating outflows and only a few self - absorbed profiles , hnc seems to be a good tracer of cold and warm gas . the hnc emission lines show , in general , broad widths ( as seen in the hco@xmath1 lines as well , figure [ fwhm - plot ] ) . this is probably because both lines are optically thick , and @xmath41v can be broader by opacity . on the other hand , hn@xmath2c is optically thin and normally shows a gaussian profile . we note that the assumption of hnc emission being optically thin produces column densities of at least one order of magnitude lower . there is a slight increase of hnc column densities with the evolution of the clumps . however , there are no differences in the hnc abundances for the different evolutionary states . the accretion of material in the star - forming clumps could explain the small rise in the hnc column densities . from the histogram with the relative abundances of n@xmath0h@xmath1 and hnc ( figure [ n2h_hnc_ratio ] ) , it can be inferred that hnc may be preferentially formed in cold gas . future chemical modeling is needed to clarify the behavior of hnc in high - mass star - forming regions . the hcn @xmath31 rotational transitions show three hyperfine components caused by the nuclear spin of the nitrogen nucleus . in the optically thin case , these three components have relative intensities of 1:3:5 . however , due to the broad linewithds found in massive star - forming regions , the components are blended . in addition , several sources exhibit spectra with extended wing emission and a wide variety of relative intensities other than those in the optically thin limit . the combination of these three factors makes it difficult to perform gaussian fits and analyze the hcn line towards the irdc clumps in this survey . for a better understanding of hcn hyperfine transitions , see the recent work of @xcite . c@xmath0h @xmath190 exhibits 6 fine and hyperfine components due to the presence of both an electron spin and a nuclear spin of the hydrogen nucleus . the spectral separation makes all of them observable in star - forming regions . however , two of them are relatively very weak and were not detected in this survey . their relative intensities in the optically thin case are given by @xcite and @xcite . c@xmath0h has been known to be a pdr tracer ( e.g. , * ? ? ? however , @xcite find that this molecule also seems to trace dense gas in early stages of star formation . they found that the distribution of c@xmath0h shows a hole around a hot core , and suggest that c@xmath0h decreases in the hot core phase . based on the results of their simple chemical modeling , they suggest that c@xmath0h may be a suitable tracer of early stages of star formation . so far , c@xmath0h has not been systematically studied in massive star - forming regions or irdcs . although we can not know the spatial distribution of c@xmath0h in the irdc clumps , we can say that there is definitely an increase in c@xmath0h detections from quiescent to active sources ( see figure [ detection_rates_per_category ] ) . the @xmath630% detection rate of this line in quiescent clumps shows that its emission is not ubiquitous in these kinds of sources as , for example , n@xmath0h@xmath1 , hnc , and hco@xmath1 are . certainly , higher angular resolution observations and mapping are needed to clarify if the c@xmath0h emission comes from the external layers of clumps that can interact with pdr emission surrounding them , or if the c@xmath0h emission comes from the dense , cold gas inside quiescent clumps . we note that c@xmath0h lines show the best @xmath41v correlation with n@xmath0h@xmath1 , suggesting that the emission originates from the same region ( see figure [ fwhm - plot ] ) . c@xmath0h column densities and abundances show no clear trend of changing with evolution . there is marginal evidence for increasing column densities with the evolution of the clumps , but it is not supported by the k - s test . hc@xmath3n is the simplest of the cyanopolyynes , molecules in the form of hc@xmath191n with n from 1 to 5 . its main progenitor , c@xmath0h@xmath0 , exists on grains mantles and is released into the gas phase during the onset of heating . thus , hc@xmath3n is associated with warm , dense gas in regions with current star formation , such as hot cores ( e.g. , * ? ? ? this may explain the high detection rates ( @xmath650% ) for this molecule in active clumps . the presence of hc@xmath3n and the rich molecular line spectra found in active cores ( see figure [ number_lines ] ) , suggest that some active clumps may have embedded hot molecular cores , as , for example , @xcite found in one irdc clump using high angular resolution observations . the hc@xmath3n linewidths are more narrow than n@xmath0h@xmath1 linewidths , and seem to correlate better with those of h@xmath2co@xmath1 and hn@xmath2c ( see figure [ fwhm - plot ] ) . column densities vary with the evolution of the clumps , increasing in more evolved regions as is shown by the median values . for hc@xmath3n abundances , no real change is clear . lccccccccc g015.05 mm1 & 45 & 0.86 & 1.82(0.15 ) & 2.39(0.85 ) & 2.64(1.27 ) & 3.20(3.52 ) & ... & 1.94(0.81 ) & ... + g015.05 mm2 & 25 & 0.35 & 2.05(0.45 ) & 1.69(0.60 ) & 7.70(5.53 ) & ... & ... & ... & ... + g015.05 mm3 & 18 & 0.33 & 1.08(0.31 ) & ... & 3.12(1.52 ) & ... & ... & ... & ... + g015.05 mm4 & 14 & 0.18 & 5.51(0.63 ) & 3.09(1.07 ) & 8.26(3.43 ) & 16.20(17.41 ) & ... & ... & ... + g015.05 mm5 & 24 & 0.32 & 1.63(0.51 ) & 3.16(1.74 ) & 7.50(4.84 ) & ... & ... & ... & ... + g015.31 mm2 & 22 & 0.43 & 2.73(2.89 ) & ... & 5.05(4.12 ) & ... & ... & ... & ... + g015.31 mm3 & 17 & 0.32 & 0.96(0.25 ) & 2.65(1.90 ) & 1.58(1.61 ) & ... & ... & ... & ... + g015.31 mm5 & 22 & 0.34 & ... & 3.22(1.83 ) & 3.98(2.79 ) & ... & ... & ... & ... + g018.82 mm2 & 77 & 0.82 & 2.74(2.75 ) & 3.91(2.38 ) & 4.21(2.55 ) & ... & ... & ... & ... + g018.82 mm3 & 30 & 0.36 & 16.75(16.99 ) & 27.86(14.46 ) & 18.52(8.67 ) & ... & ... & ... & ... + g018.82 mm4 & 43 & 1.26 & 0.85(0.14 ) & 6.73(5.33 ) & 3.98(2.25 ) & ... & ... & 0.77(0.83 ) & ... + g018.82 mm6 & 46 & 0.88 & 1.39(0.14 ) & 0.60(0.21 ) & 0.91(0.40 ) & ... & ... & ... & ... + g019.27 mm2 & 46 & 0.65 & 3.36(0.22 ) & ... & 6.30(2.15 ) & 6.40(7.08 ) & ... & ... & ... + g022.35 mm1 & 93 & 2.19 & 0.19(0.05 ) & 0.39(0.15 ) & 0.51(0.21 ) & ... & ... & ... & ... + g022.35 mm2 & 32 & 0.29 & ... & ... & 6.94(3.70 ) & ... & ... & ... & ... + g023.60 mm7 & 49 & 0.43 & 9.45(9.31 ) & ... & 8.00(3.82 ) & ... & ... & ... & ... + g023.60 mm9 & 32 & 0.41 & 1.67(0.27 ) & ... & 3.58(1.72 ) & 3.68(5.28 ) & ... & ... & ... + g024.08 mm2 & 48 & 0.72 & 0.98(0.16 ) & 2.49(1.01 ) & 3.52(1.57 ) & ... & ... & ... & ... + g024.08 mm3 & 33 & 0.48 & 1.94(1.92 ) & 3.72(1.85 ) & 3.98(1.77 ) & ... & ... & ... & ... + g024.08 mm4 & 45 & 0.63 & 0.76(0.21 ) & 3.86(3.50 ) & 2.92(1.49 ) & ... & ... & ... & ... + g024.33 mm2 & 44 & 0.56 & 5.44(0.32 ) & 2.43(0.63 ) & 10.94(2.07 ) & ... & ... & 7.58(1.95 ) & ... + g024.33 mm3 & 53 & 0.70 & 3.77(0.24 ) & 1.97(0.50 ) & 1.22(0.41 ) & 4.85(6.65 ) & ... & 4.01(1.28 ) & 5.64(2.02 ) + g024.33 mm4 & 53 & 0.72 & 2.19(0.19 ) & 1.10(0.42 ) & 1.10(0.40 ) & ... & ... & ... & ... + g024.33 mm5 & 51 & 0.67 & 3.24(0.22 ) & 4.18(0.68 ) & 1.18(0.42 ) & 3.92(5.95 ) & ... & 8.10(1.39 ) & ... + g024.33 mm7 & 26 & 0.49 & 1.89(0.17 ) & 2.49(0.47 ) & 2.15(0.78 ) & 1.24(2.46 ) & ... & ... & 2.78(1.51 ) + g024.33 mm8 & 42 & 0.55 & ... & 2.50(0.95 ) & 4.69(1.80 ) & ... & ... & ... & ... + g024.33 mm9 & 43 & 0.31 & ... & 7.24(2.81 ) & 10.96(4.29 ) & ... & ... & ... & ... + g024.33 mm11 & 32 & 0.55 & 2.76(0.17 ) & 5.16(1.14 ) & 7.34(2.73 ) & ... & ... & ... & ... + g024.60 mm2 & 78 & 1.54 & 0.81(0.06 ) & 0.98(0.32 ) & 0.96(0.41 ) & 0.75(0.26 ) & 2.25(0.80 ) & ... & ... + g025.04 mm2 & 59 & 0.91 & 0.68(0.12 ) & 1.39(0.56 ) & 1.87(0.80 ) & ... & ... & ... & ... + g025.04 mm4 & 62 & 0.89 & 2.74(0.12 ) & 1.69(0.61 ) & 4.82(1.13 ) & 1.22(2.49 ) & ... & 3.50(0.73 ) & ... + g027.75 mm2 & 17 & 0.32 & 3.12(3.15 ) & ... & ... & ... & ... & ... & ... + g027.94 mm1 & 41 & 0.37 & 4.44(4.46 ) & ... & 5.28(1.83 ) & 5.19(8.39 ) & ... & ... & ... + g028.04 mm1 & 44 & 0.47 & 4.87(4.79 ) & 4.80(2.04 ) & 11.79(5.69 ) & 2.73(4.87 ) & 8.07(2.62 ) & 6.17(1.80 ) & ... + g028.08 mm1 & 29 & 0.61 & 1.03(1.02 ) & ... & 3.14(1.66 ) & ... & ... & ... & ... + g028.23 mm1 & 59 & 1.23 & 2.54(2.55 ) & 0.81(0.27 ) & 0.62(0.22 ) & ... & ... & ... & ... + g028.28 mm4 & 43 & 0.51 & 2.96(0.26 ) & 3.55(0.73 ) & 3.20(1.17 ) & 5.19(5.99 ) & ... & ... & ... + g028.37 mm1 & 318 & 4.02 & 1.14(0.04 ) & 0.62(0.10 ) & 1.42(0.24 ) & 1.07(0.60 ) & 3.60(0.36 ) & 2.48(0.22 ) & 3.98(0.34 ) + g028.37 mm2 & 118 & 1.23 & 2.04(0.15 ) & 1.97(0.36 ) & 5.18(1.43 ) & 3.70(5.25 ) & 3.96(1.73 ) & 2.94(1.22 ) & ... + g028.37 mm4 & 115 & 1.51 & 3.10(0.12 ) & 1.92(0.20 ) & 2.89(0.31 ) & 4.17(3.73 ) & 5.65(0.71 ) & 5.33(0.42 ) & 9.20(0.65 ) + g028.37 mm6 & 87 & 1.71 & 1.56(0.06 ) & 1.59(0.30 ) & 1.68(0.27 ) & 1.06(1.38 ) & 4.48(0.71 ) & 2.51(0.30 ) & 5.96(0.52 ) + g028.37 mm9 & 67 & 1.21 & 1.26(0.09 ) & 1.64(0.29 ) & 1.26(0.24 ) & ... & ... & ... & ... + g028.37 mm11 & 53 & 0.60 & 1.74(0.34 ) & 1.36(0.47 ) & 11.93(6.30 ) & 4.36(6.98 ) & ... & ... & ... + g028.37 mm12 & 51 & 0.97 & 0.92(0.93 ) & 2.49(1.42 ) & 3.71(2.94 ) & ... & 3.14(1.58 ) & ... & ... + g028.37 mm13 & 20 & 0.28 & 5.16(5.59 ) & ... & ... & ... & ... & ... & ... + g028.53 mm3 & 85 & 1.78 & 1.39(1.37 ) & 1.19(0.36 ) & 1.34(0.58 ) & ... & ... & ... & ... + g028.53 mm5 & 44 & 0.63 & 4.63(4.83 ) & 2.39(1.62 ) & 5.48(3.21 ) & ... & ... & ... & ... + g028.53 mm7 & 44 & 0.86 & 1.28(0.14 ) & 1.06(0.31 ) & 4.03(1.41 ) & ... & ... & ... & ... + g028.53 mm8 & 35 & 0.54 & 1.84(0.26 ) & 6.53(5.13 ) & 6.35(3.27 ) & 3.73(2.12 ) & ... & ... & ... + g028.53 mm9 & 61 & 0.94 & 2.00(0.16 ) & 2.53(1.01 ) & 3.27(1.07 ) & ... & ... & ... & ... + g028.53 mm10 & 90 & 1.28 & 1.59(0.15 ) & 2.22(0.85 ) & 3.83(1.14 ) & 3.96(5.40 ) & ... & ... & ... + g028.67 mm1 & 27 & 0.27 & ... & 6.07(3.02 ) & 10.93(5.20 ) & ... & ... & ... & ... + g028.67 mm2 & 38 & 0.76 & 0.95(0.14 ) & ... & 3.51(2.09 ) & ... & ... & ... & ... + g030.14 mm1 & 40 & 0.64 & 2.12(2.33 ) & 1.83(1.05 ) & 2.50(1.28 ) & ... & ... & ... & ... + g030.57 mm1 & 117 & 1.40 & 3.16(0.11 ) & ... & 2.75(0.87 ) & 1.89(2.95 ) & 6.10(1.11 ) & 2.04(0.70 ) & ... + g030.57 mm3 & 17 & 0.34 & 2.92(0.55 ) & ... & 18.84(10.70 ) & 7.12(10.94 ) & ... & ... & ... + g030.97 mm1 & 98 & 1.08 & 3.38(0.14 ) & 4.80(0.46 ) & 3.75(0.62 ) & ... & 3.43(1.13 ) & ... & ... + g031.97 mm5 & 29 & 0.35 & 6.77(0.39 ) & 3.63(0.85 ) & 10.27(3.34 ) & 8.52(11.04 ) & ... & ... & ... + g031.97 mm7 & 36 & 0.59 & 2.89(0.16 ) & 2.05(0.40 ) & 5.13(4.79 ) & 3.66(5.80 ) & ... & 1.70(0.96 ) & ... + g031.97 mm8 & 38 & 0.31 & 10.84(0.62 ) & ... & 10.76(3.73 ) & ... & ... & ... & ... + g033.69 mm1 & 97 & 1.71 & 2.08(0.06 ) & 1.00(0.15 ) & 1.72(0.20 ) & ... & 5.61(0.75 ) & 2.66(0.35 ) & 3.09(0.60 ) + g033.69 mm2 & 75 & 0.73 & 4.34(0.29 ) & 6.17(2.51 ) & 6.11(1.56 ) & 4.84(0.67 ) & 13.77(2.24 ) & 5.87(1.66 ) & ... + g033.69 mm3 & 52 & 0.48 & 4.97(0.58 ) & 7.48(3.12 ) & 10.73(4.33 ) & ... & ... & ... & ... + g033.69 mm4 & 65 & 1.00 & 3.06(0.21 ) & 1.77(0.65 ) & 3.38(0.83 ) & ... & 5.59(1.97 ) & 3.09(1.02 ) & ... + g033.69 mm5 & 44 & 0.49 & 4.42(0.50 ) & 7.90(3.93 ) & 8.69(3.61 ) & 7.70(6.26 ) & ... & ... & ... + g033.69 mm11 & 37 & 0.65 & ... & 3.88(1.46 ) & 6.14(2.33 ) & 2.28(4.42 ) & ... & ... & ... + g034.43 mm1 & 548 & 5.85 & 2.34(0.04 ) & 0.82(0.09 ) & 2.33(0.27 ) & 1.32(0.81 ) & 5.70(0.36 ) & 1.47(0.25 ) & 5.94(0.36 ) + g034.43 mm5 & 89 & 1.75 & 1.70(0.05 ) & 1.10(0.15 ) & 2.99(0.45 ) & ... & 3.07(0.71 ) & ... & ... + g034.43 mm7 & 59 & 0.85 & 1.31(0.12 ) & 1.26(0.39 ) & 2.69(0.70 ) & 2.41(3.42 ) & 2.43(1.38 ) & ... & ... + g034.43 mm8 & 69 & 0.64 & 2.39(0.26 ) & 4.23(1.89 ) & 3.98(1.58 ) & 5.56(9.33 ) & ... & ... & ... + g034.77 mm1 & 111 & 1.11 & 3.68(3.65 ) & 4.08(0.71 ) & 2.20(0.72 ) & ... & 4.21(1.17 ) & ... & ... + g034.77 mm3 & 18 & 0.25 & 3.55(0.60 ) & 2.03(0.71 ) & 11.20(4.72 ) & 12.12(20.78 ) & ... & ... & ... + g035.39 mm7 & 50 & 0.61 & 3.92(0.27 ) & 9.77(1.54 ) & 6.21(1.34 ) & 7.13(6.73 ) & 5.27(2.41 ) & ... & ... + g035.59 mm1 & 22 & 0.25 & 11.25(12.28 ) & 17.36(8.70 ) & 7.02(4.36 ) & ... & ... & ... & ... + g035.59 mm2 & 19 & 0.35 & 2.03(0.29 ) & 5.09(1.59 ) & 3.05(1.14 ) & ... & ... & ... & ... + g035.59 mm3 & 30 & 0.42 & 2.40(0.55 ) & 2.69(0.70 ) & 3.00(1.13 ) & 4.37(5.26 ) & 3.32(2.99 ) & ... & ... + g036.67 mm1 & 38 & 0.90 & 1.10(1.13 ) & 1.75(1.49 ) & 1.13(0.66 ) & 1.75(1.74 ) & ... & ... & ... + g036.67 mm2 & 38 & 1.22 & ... & 0.54(0.24 ) & 0.95(0.39 ) & ... & ... & ... & ... + g038.95 mm1 & 104 & 3.33 & 0.40(0.02 ) & 1.28(0.20 ) & 0.51(0.15 ) & 0.72(0.58 ) & 1.41(0.58 ) & 0.27(0.16 ) & ... + g038.95 mm2 & 82 & 0.82 & 2.10(0.25 ) & 5.66(1.29 ) & 2.99(1.12 ) & ... & 4.23(2.10 ) & ... & ... + g038.95 mm3 & 57 & 0.88 & 3.40(3.38 ) & 3.88(1.00 ) & 2.69(1.06 ) & 3.20(0.46 ) & ... & ... & ... + g038.95 mm4 & 44 & 0.40 & 4.89(4.95 ) & 4.54(1.69 ) & 5.81(2.28 ) & 5.36(0.97 ) & ... & ... & 6.16(4.48 ) + g053.11 mm1 & 227 & 2.43 & 1.28(0.10 ) & 1.95(0.20 ) & 0.81(0.24 ) & ... & 3.28(0.57 ) & ... & ... + g053.11 mm2 & 55 & 0.63 & 4.39(4.42 ) & 4.47(0.90 ) & 2.97(0.85 ) & 2.58(0.98 ) & ... & ... & ... + g053.11 mm4 & 34 & 0.34 & 3.29(0.50 ) & 11.95(1.79 ) & 4.29(1.14 ) & 2.79(1.06 ) & 4.69(3.59 ) & ... & ... + g053.11 mm5 & 25 & 0.22 & 13.32(13.61 ) & 5.79(2.17 ) & 7.07(2.69 ) & 4.97(1.14 ) & ... & ... & ... + g053.25 mm1 & 52 & 0.61 & 2.36(0.24 ) & 2.28(0.75 ) & 1.98(0.68 ) & 2.54(3.40 ) & ... & ... & ... + g053.25 mm3 & 17 & 0.20 & 6.87(7.23 ) & 7.48(3.09 ) & 6.77(2.63 ) & 15.56(26.01 ) & ... & ... & ... + g053.25 mm4 & 37 & 0.40 & 3.99(0.37 ) & 6.01(1.37 ) & 3.04(1.07 ) & ... & ... & ... & ... + g053.25 mm5 & 17 & 0.20 & ... & 12.49(4.32 ) & 6.76(2.55 ) & 10.04(22.70 ) & ... & ... & ... + g053.25 mm6 & 34 & 0.30 & 13.05(13.20 ) & 8.97(2.12 ) & 5.61(2.10 ) & ... & ... & ... & ... + g053.31 mm2 & 34 & 0.59 & 2.09(2.07 ) & 1.13(0.42 ) & 1.69(0.65 ) & ... & ... & ... & ... + hnco is a high density tracer that has been found in regions that span a large range of temperatures ( 10 - 500 k ; * ? ? ? * ; * ? ? ? * ; * ? ? ? the formation of hnco is inefficient if only gas - phase reactions are considered @xcite . in hot cores , the gas formation chemistry can explain the observed hnco abundances when reactions on the surfaces of grains are included . hnco is efficiently formed on grain mantles and processed in more complex molecules . however , hnco is not directly ejected from the dust . instead , it is formed by the dissociation of more complex molecules ( formed by hnco on the grain mantles ) once they are released to the gas phase @xcite . @xcite found evidence that hnco and sio may have a common production mechanism , presumably based on shock chemistry . @xcite found hnco lines that exhibit the same characteristics as other well - known shock tracers such as ch@xmath3oh and some sulfur - bearing species ( so and so@xmath0 ) . these authors suggest that hnco is a shock tracer which is explained by a combination of grain surface and gas phase chemistry . hnco abundances would rise in shock regions because this molecule can be directly ejected to the gas phase through grain sputtering and because the efficiency of neutral - neutral reactions increases in the gas phase due to higher gas temperatures . in our irdc sample , hnco emission is detected slightly more often in regions with signs of star formation . the hnco line widths seem to correlate with those of n@xmath0h@xmath1 . hnco profiles show no evident signatures of being a tracer of shocks in most of the sources , except in g028.37 mm1 and g028.37 mm4 ( which represent 10% of the sources with hnco detection ) . in these two sources , the hnco spectrum presents a blue wing which is also observed in sio . hnco column densities rise with the evolution of the clumps , as is shown by the median values . however , hnco abundances do not show this trend . it is well - known that sio is a powerful tracer of molecular gas associated with shocks @xcite . its abundance is highly enhanced in molecular outflows , with respect to the ambient abundances . in some cases these enhancements can be up to @xmath123 @xcite . the increase of the sio abundance is due to the silicate grain destruction , giving rise to the injection into gas of si atoms and/or si - bearing species , and the subsequent high temperature gas - phase chemistry @xcite . the previous sio production mechanism explains spectral lines that show extended wing emission with broad line widths caused by the interaction of outflows and the surrounding medium . however , an sio component coming from quiescent gas is not well understood . @xcite detected extended narrow sio emission not associated with signs of star formation in an irdc . they suggest that this sio component could be produced by the following processes : remnants of large - scale shocks caused by the formation process of the irdc , decelerated shocked gas associated with large scale outflows from neighboring massive protostars , recently processed material associated with the youngest jets / outflows , and/or an undetected and widespread lower mass protostar population . in our irdc sample , sio emission is mostly detected in clumps with signs of star formation . the only exception is the quiescent clump g024.33 mm7 which shows a narrow line profile ( 1.4 ) and the lowest sio abundance in the sample . the average sio line width in the sample is @xmath65 . half of the sources with sio detection show clear extended wing emission indicating the presence of outflow activity . we have carried out a multi - line survey at 3 mm toward 37 irdcs , containing 159 clumps , in order to investigate the behavior of the different molecular tracers and search for chemical variations through an evolutionary sequence based on _ spitzer _ irac and mips emission . we observed n@xmath0h@xmath1 , hnc , hn@xmath2c , hco@xmath1 , h@xmath2co@xmath1 , hcn , c@xmath0h , hc@xmath3n , hnco , and sio lines with the mopra 22 m telescope located in australia . after eliminating clumps that are not located in irdcs and pairs of clumps that are placed within one mopra beam , we base our study on 92 sources . hnc and n@xmath0h@xmath1 lines are detected in almost every irdc clumps at every evolutionary stage , indicating that their presence does not depend on the star formation activity . on the other hand , hc@xmath3n , hnco , and sio lines are predominantly detected in later stages of evolution , as expected from their formation paths . the line widths of n@xmath0h@xmath1 slightly increase with the evolution of the clumps , which is likely produced by the rise of turbulence due to the enhancement of the star formation activity at later evolutionary stages . the increase is modest because , due to the large mopra beam , we are also tracing the bulk motions of the gas , instead of just the densest regions associated with star formation . optical depth calculations show that the n@xmath0h@xmath1 line is mostly optically thin ( median of 0.8 ) and the c@xmath0h line is moderately optically thick ( median of 2.4 ) . hco@xmath1 and hnc lines are optically thick ( medians of 21 and 19 , respectively ) , while their isotopologues are optically thin ( median of 0.4 for both ) . n@xmath0h@xmath1 opacities show no variations with the evolution of the clumps , whereas c@xmath0h , hco@xmath1 and hnc show a slight decrease with the rise of star formation activity . in general , column densities of the different molecules change for the different evolutionary stages defined by @xcite and increase with the evolution of the clumps , with the exception of c@xmath0h . however , this is not generally true for molecular abundances ( i.e. , after dividing by the total h@xmath0 column density inferred from 1.2 mm continuum emission ) . only the increases of n@xmath0h@xmath1 and hco@xmath1 abundances are statistically significant and reflect chemical evolution . this is consistent with the results of @xcite , who included both molecules in their chemical modeling of a massive star - forming region . although it is expected a rise of the hco@xmath1 abundance with the evolution of the clumps , it is not clear why n@xmath0h@xmath1 also follows this trend . the n@xmath0h@xmath98/hco@xmath1 abundance ratio acts as a chemical clock , increasing its value from intermediate to active and red clumps . this observed trend is consistent with the theoretical predictions . the chemical models suggest that when clumps warm up , they release co from grain mantles to the gas phase . this rise of co increases the amount of hco@xmath1 , because co is its main supplier , with respect to n@xmath0h@xmath1 , because co is its main destroyer . however , the observed trend does not extend to quiescent clumps . this could be due to observational limitations or because the theoretical predictions are incorrect at very early stages of evolution . we also find that the n@xmath0h@xmath98/hnc abundance ratio increases with the evolution of the clumps , from quiescent to red clumps . it is not clear why this ratio behaves in this way , but it suggests that hnc may be preferentially formed in cold gas . j.m.j . gratefully acknowledges funding support from nsf grant no . g.g . acknowledges support from conicyt projects fondap no . 15010003 and basal pfb-06 . we also thank the anonymous referee for helpful comments . battersby , c. , bally , j. , jackson , j. m. , et al . 2010 , , 721 , 222 beltrn , m. t. , cesaroni , r. , neri , r. , et al . 2005 , , 435 , 901 beuther , h. , & sridharan , t. k. 2007 , , 668 , 348 beuther , h. , semenov , d. , henning , t. , & linz , h. 2008 , , 675 , l33 bisschop , s. e. , jrgensen , j. k. , van dishoeck , e. f. , & de wachter , e. b. m. 2007 , , 465 , 913 blackman , g. l. , brown , r. d. , godfrey , p. d. , & gunn , h. i. 1976 , , 261 , 395 blake , g. a. , sutton , e. c. , masson , c. r. , & phillips , t. g. 1987 , , 315 , 621 botschwina , p. 1984 , chemical physics letters , 107 , 535 botschwina , p. , horn , m. , flugge , j. , & seeger , s. 1993 , j. chem . faraday trans . , 623 , 2219 busquet , g. , estalella , r. , zhang , q. , et al . 2011 , , 525 , a141 carey , s. j. , clark , f. o. , egan , m. p. , et al . 1998 , , 508 , 721 carey , s. j. , feldman , p. a. , redman , r. o. , et al . 2000 , , 543 , l157 caselli , p. , myers , p. c. , & thaddeus , p. 1995 , , 455 , l77 caselli , p. , hartquist , t. w. , & havnes , o. 1997 , , 322 , 296 caselli , p. , benson , p. j. , myers , p. c. , & tafalla , m. 2002 , , 572 , 238 chambers , e. t. , jackson , j. m. , rathborne , j. m. , & simon , r. 2009 , , 181 , 360 chapman , j. f. , millar , t. j. , wardle , m. , burton , m. g. , & walsh , a. j. 2009 , , 394 , 221 chen , x. , shen , z .- q . , li , j .- j . , xu , y. , & he , j .- h . 2010 , , 710 , 150 chen , h .- r . , liu , s .- y . , su , y .- , & wang , m .- y . 2011 , , 743 , 196 clemens , d. p. 1985 , , 295 , 422 crapsi , a. , caselli , p. , walmsley , c. m. , myers , p. c. , tafalla , m. , lee , c. w. , & bourke , t. l. 2005 , , 619 , 379 cyganowski , c. j. , whitney , b. a. , holden , e. , et al . 2008 , , 136 , 2391 cyganowski , c. j. , brogan , c. l. , hunter , t. r. , churchwell , e. , & zhang , q. 2011 , , 729 , 124 daniel , f. , cernicharo , j. , & dubernet , m .- 2006 , , 648 , 461 devine , k. e. , chandler , c. j. , brogan , c. , et al . 2011 , , 733 , 44 egan , m. p. , shipman , r. f. , price , s. d. , et al . 1998 , , 494 , l199 emprechtinger , m. , caselli , p. , volgenau , n. h. , stutzki , j. , & wiedner , m. c. 2009 , , 493 , 89 fandez , s. , bronfman , l. , garay , g. , et al . 2004 , , 426 , 97 fontani , f. , palau , a. , caselli , p. , et al . 2011 , , 529 , l7 foster , j. b. , jackson , j. m. , barnes , p. j. , et al . 2011 , , 197 , 25 fuente , a. , martin - pintado , j. , cernicharo , j. , & bachiller , r. 1993 , , 276 , 473 fuller , g. a. , williams , s. j. , & sridharan , t. k. 2005 , , 442 , 949 garden , r. p. , hayashi , m. , hasegawa , t. , gatley , i. , & kaifu , n. 1991 , , 374 , 540 havenith , m. , zwart , e. , leo meerts , w. , & ter meulen , j. j. 1990 , , 93 , 8446 jackson , j. m. , armstrong , j. t. , & barrett , a. h. 1984 , , 280 , 608 jackson , j. m. , rathborne , j. m. , shah , r. y. , et al . 2006 , , 163 , 145 jackson , j. m. , finn , s. c. , rathborne , j. m. , chambers , e. t. , & simon , r. 2008 , , 680 , 349 jimnez - serra , i. , caselli , p. , tan , j. c. , et al . 2010 , , 406 , 187 jrgensen , j. k. , schier , f. l. , & van dishoeck , e. f. 2004 , , 416 , 603 kauffmann , j. , & pillai , t. 2010 , , 723 , l7 kim , g. , lee , c. w. , kim , j. , et al . 2010 , journal of korean astronomical society , 43 , 9 kim , s .- j . , kim , h .- lee , y. , minh , y. c. , balasubramanyam , r. , burton , m. g. , millar , t. j. , & lee , d .- w . 2006 , , 162 , 161 lapinov , a. v. , golubiatnikov , g. y. , markov , v. n. , & guarnieri , a. 2007 , astronomy letters , 33 , 121 lattanzi , v. , walters , a. , drouin , b. j. , & pearson , j. c. 2007 , , 662 , 771 lee , j .- e . , bergin , e. a. , & evans , n. j. , ii 2004 , , 617 , 360 lo , n. , et al . 2009 , , 395 , 1021 loughnane , r. m. , redman , m. p. , thompson , m. a. , et al . 2012 , , 420 , 1367 machin , l. , & roueff , e. 2006 , , 460 , 953 markwardt , c. b. 2009 , astronomical data analysis software and systems xviii , 411 , 251 martin - pintado , j. , bachiller , r. , & fuente , a. 1992 , , 254 , 315 miettinen , o. , harju , j. , haikala , l. k. , & pomrn , c. 2006 , , 460 , 721 miettinen , o. , hennemann , m. , & linz , h. 2011 , , 534 , a134 mollaaghababa , r. , gottlieb , c. a. , vrtilek , j. m. , & thaddeus , p. 1991 , , 368 , l19 mller , h. s. p. , thorwirth , s. , roth , d. a. , & winnewisser , g. 2001 , , 370 , l49 mller , h. s. p. , schlder , f. , stutzki , j. , & winnewisser , g. 2005 , journal of molecular structure , 742 , 215 berg , k. i. , van broekhuizen , f. , fraser , h. j. , et al . 2005 , , 621 , l33 ossenkopf , v. , & henning , t. 1994 , , 291 , 943 padovani , m. , walmsley , c. m. , tafalla , m. , galli , d. , mller , h. s. p. 2009 , , 505 , 1199 pagani , l. , daniel , f. , & dubernet , m .- 2009 , , 494 , 719 perault , m. , omont , a. , simon , g. , et al . 1996 , , 315 , l165 peretto , n. , & fuller , g. a. 2009 , , 505 , 405 pillai , t. , wyrowski , f. , carey , s. j. , & menten , k. m. 2006 , , 450 , 569 pillai , t. , kauffmann , j. , wyrowski , f. , et al . 2011 , , 530 , a118 pirogov , l. , zinchenko , i. , caselli , p. , johansson , l. e. b. , & myers , p. c. 2003 , , 405 , 639 prasad , s. s. , & huntress , w. t. , jr . 1980 , , 239 , 151 purcell , c. r. , et al . 2006 , , 367 , 553 ragan , s. e. , bergin , e. a. , & wilner , d. 2011 , , 736 , 163 rathborne , j. m. , jackson , j. m. , & simon , r. 2006 , , 641 , 389 rathborne , j. m. , simon , r. , & jackson , j. m. 2007 , , 662 , 1082 rathborne , j. m. , jackson , j. m. , zhang , q. , & simon , r. 2008 , , 689 , 1141 rathborne , j. m. , jackson , j. m. , chambers , e. t. , stojimirovic , i. , simon , r. , shipman , r. , & frieswijk , w. 2010 , , 715 , 310 raymonda , j. w. , muenter , j. s. , & klemperer , w. a. 1970 , , 52 , 3458 rodrguez - fernndez , n. j. , tafalla , m. , gueth , f. , & bachiller , r. 2010 , , 516 , a98 rodriguez - franco , a. , martin - pintado , j. , & fuente , a. 1998 , , 329 , 1097 rygl , k. l. j. , wyrowski , f. , schuller , f. , & menten , k. m. 2010 , , 515 , a42 savage , c. , apponi , a. j. , ziurys , l. m. , & wyckoff , s. 2002 , , 578 , 211 sakai , t. , sakai , n. , kamegai , k. , hirota , t. , yamaguchi , n. , shiba , s. , & yamamoto , s. 2008 , , 678 , 1049 sakai , t. , sakai , n. , hirota , t. , & yamamoto , s. 2010 , , 714 , 1658 sakai , t. , sakai , n , furuya , k. , aikawa , y. , hirota , t. , yamamoto , s. , 2012 , apj , in press . sanhueza , p. , garay , g. , bronfman , l. , et al . 2010 , , 715 , 18 schilke , p. , walmsley , c. m. , pineau des forets , g. , & flower , d. r. 1997 , , 321 , 293 schier , f. l. , van der tak , f. f. s. , van dishoeck , e. f. , & black , j. h. 2005 , , 432 , 369 simon , r. , jackson , j. m. , rathborne , j. m. , & chambers , e. t. 2006 , , 639 , 227 simon , r. , rathborne , j. m. , shah , r. y. , jackson , j. m. , & chambers , e. t. 2006 , , 653 , 1325 snyder , l. e. , hollis , j. m. , & watson , w. d. 1977 , , 212 , 79 thorwirth , s. , mller , h. s. p. , & winnewisser , g. 2000 , journal of molecular spectroscopy , 204 , 133 tideswell , d. m. , fuller , g. a. , millar , t. j. , & markwick , a. j. 2010 , , 510 , a85 tucker , k. d. , kutner , m. l. , & thaddeus , p. 1974 , , 193 , l115 turner , b. e. , & thaddeus , p. 1977 , , 211 , 755 van der tak , f. f. s. , mller , h. s. p. , harding , m. e. , & gauss , j. 2009 , , 507 , 347 vasyunina , t. , linz , h. , henning , t. , zinchenko , i. , beuther , h. , & voronkov , m. 2011 , , 527 , a88 viti , s. , collings , m. p. , dever , j. w. , mccoustra , m. r. s. , & williams , d. a. 2004 , , 354 , 1141 wang , y. , zhang , q. , rathborne , j. m. , jackson , j. , & wu , y. 2006 , , 651 , l125 yamaguchi , y. , richards , c. a. , jr . , & schaefer , h. f. , iii 1994 , , 101 , 8945 zhang , q. , wang , y. , pillai , t. , & rathborne , j. 2009 , , 696 , 268 zhang , q. , & wang , k. 2011 , , 733 , 26 zinchenko , i. , henkel , c. , & mao , r. q. 2000 , , 361 , 1079	we have observed 37 infrared dark clouds ( irdcs ) , containing a total of 159 clumps , in high - density molecular tracers at 3 mm using the 22-meter atnf mopra telescope located in australia . after determining kinematic distances , we eliminated clumps that are not located in irdcs and clumps with a separation between them of less than one mopra beam . our final sample consists of 92 irdc clumps . the most commonly detected molecular lines are ( detection rates higher than 8% ) : n@xmath0h@xmath1 , hnc , hn@xmath2c , hco@xmath1 , h@xmath2co@xmath1 , hcn , c@xmath0h , hc@xmath3n , hnco , and sio . we investigate the behavior of the different molecular tracers and look for chemical variations as a function of an evolutionary sequence based on _ spitzer _ irac and mips emission . we find that the molecular tracers behave differently through the evolutionary sequence and some of them can be used to yield useful relative age information . the presence of hnc and n@xmath0h@xmath1 lines do not depend on the star formation activity . on the other hand , hc@xmath3n , hnco , and sio are predominantly detected in later stages of evolution . optical depth calculations show that in irdc clumps the n@xmath0h@xmath1 line is optically thin , the c@xmath0h line is moderately optically thick , and hnc and hco@xmath1 are optically thick . the hcn hyperfine transitions are blended , and , in addition , show self - absorbed line profiles and extended wing emission . these factors combined prevent the use of hcn hyperfine transitions for the calculation of physical parameters . total column densities of the different molecules , except c@xmath0h , increase with the evolutionary stage of the clumps . molecular abundances increase with the evolutionary stage for n@xmath0h@xmath1 and hco@xmath1 . the n@xmath0h@xmath1/hco@xmath1 and n@xmath0h@xmath1/hnc abudance ratios act as chemical clocks , increasing with the evolution of the clumps .
red supergiants ( rsgs ) are an evolved he - burning phase in the life of a moderately massive star ( 10 - 25 m@xmath5 ) . their location in the h - r diagram has , in the past , been at odds with the predictions of stellar evolutionary theory , occupying a position that was too cold and too luminous to agree with the position of the evolutionary tracks . levesque et al . ( 2005 , 2006 ; hereafter papers i and ii ) and massey et al . ( 2009 ) have used the newest generation of the marcs stellar atmosphere models ( gustafsson et al . 1975 , 2003 , 2008 ; plez et al . 1992 ; plez 2003 ) to fit model spectral energy distributions to moderate - resolution spectrophotometry of red supergiants in the milky way , magellanic clouds , and m31 . the new physical parameters derived from this fitting shifted the properties of rsgs to higher effective temperatures and lower luminosities , bringing them into generally excellent agreement with the evolutionary tracks . however , the rsgs includes several oddball members that require a more detailed analysis . one classic example is the case of vy cma , a well - studied dust - enshrouded rsg in the milky way . massey et al . ( 2006 ) fit a spectrum of vy cma with the marcs models and derived an effective temperature ( t@xmath6 ) that was much warmer than previous estimates ( le sidaner & le bertre 1996 ) and brought this previously troublesome star into much better agreement with the evolutionary tracks ; however , determination of vy cma s stellar luminosity remains difficult due to the challenges posed by the large structured dust - reflection nebula that surrounds the star ( monnier et al . 1999 ; smith et al . 2001 ; smith 2004 , humphreys et al . 2005 ; massey et al . 2008 ) and its uncertain distance ( choi et al . another example is the unusual magellanic cloud rsgs with variable effective temperatures and dust production as described in levesque et al . ( 2007 ) and massey et al . ( 2007 ) , and henceforth referred to as levesque - massey variables . these stars show considerable changes in their effective temperatures and @xmath7 magnitudes and a common set of variations - when they are at their warmest they are also brighter , dustier , and more luminous . they are recognizable in their cooler states by occupying the forbidden " region on the h - r diagram to the right of the hayashi track , where stars are no longer expected to be in hydrostatic equilibrium ( hayashi & hoshi 1961 ) . they are currently believed to be in an unstable ( and short - lived ) evolutionary phase not previously observed in rsgs . woh g64 ( @xmath804:55:10.49 , @xmath9=-68:20:29.8 , from buchanan et al . 2006 ) is another unusual red supergiant in the large magellanic cloud ( lmc ) , originally discovered by westerlund et al . ( 1981 ) , with a number of properties that set it apart from the rest of the lmc rsg population . it is an iras source ( iras 04553 - 6825 ) , and is surrounded by an optically thick dust torus ( elias et al . 1986 , roche et al . 1993 , ohnaka et al . studies of the optical spectra by elias et al . ( 1986 ) and van loon et al . ( 2005 ) show a spectrum dominated by very strong tio bands , which has led to assignments of spectral types as late as m7 - 8 , by far the latest spectral type assigned to an lmc rsg ( cf . , paper ii ) , with corresponding extreme physical properties , which would make woh g64 the star with the largest known radius , and the rsg with the largest known luminosity and mass ( van loon et al . it is also a known source of oh , sio , and h@xmath10o masers ( wood et al . 1986 , van loon et al . 1996 , 1998 , 2001 ) , with two different maser components , suggesting a substantial mass outflow and two different expanding dust shells ( marshall et al . 2004 ) . finally , elias et al . ( 1986 ) report detections of h@xmath11 , [ oi ] , and [ nii ] emission in the optical spectrum . woh g64 is neither isolated nor in a well populated region . the star is not in any of the luck - hodge ob associations ( lucke & hodge 1970 ) . inspection of the ubvr ccd images of the lmc described by massey ( 2002 ) reveals that although there are bright neighboring blue stars in the vicinity ( such as [ m2002 ] 22861 , located 70 `` away , or the star bi 23 , classified as b0 ii by conti et al . 1986 , located 110 '' away ) , the region is not rich in bright , blue stars . we have investigated this unique rsg in more detail , applying a similar analysis used in papers i and ii to examine the physical properties that can be derived from the optical spectrum . we obtain moderate - resolution spectrophotometry of woh g64 using imacs on the baade 6.5 m magellan telescope at las campanas observatory ( [ sec : obs ] ) , confirm this star s membership in the lmc ( [ sec : mem ] ) , and compare the parameters that we determine from our model fitting to those presented in past work and predictions of evolutionary theory ( [ sec : prop ] ) . finally , we discuss the intriguing issue of emission line features in woh g64 ( [ sec : elines ] ) and summarize our findings on this unusual star ( [ sec : disc ] ) . a flux - calibrated spectrum of woh g64 was obtained on ut 2008 december 10 using the imacs spectrometer on the baade 6.5-m magellan telescope during an observing run primarily devoted to an unrelated project . a 300 line mm@xmath4 grating was used in first order for spectral coverage from 3500 to 9500 . a 0.9``-wide long slit was used , resulting in a spectral resolution of 4.3 . the observation consisted of 3x600 sec exposures , and were obtained under photometric conditions . the seeing , measured in the visible off the guide camera ( which is off - axis ) registered 0.5 - 0.6 '' during the exposure . we measure 0.7 " full width half maximum intensity on our spectra . the airmass was approximately 1.5 at the time . the slit was oriented to the parallactic angle ; we note that imacs also includes an atmospheric dispersion compensator , so our relative fluxes from the red to the blue should be good despite the modestly large airmass . flat - fielding was performed by means of a series of flats obtained on a projector screen ; wavelength calibration ( he - ne - ar lamps ) were obtained immediately after the three exposures . a large number of bias frames ( 60 ) were averaged and used to remove residual two dimensional structure after overscan correction and trimming . in addition to the observation of woh g64 , observations of three lmc rsgs with known radial velocities from massey & olsen ( 2003 ) were obtained immediately prior to the woh g64 observation in order to check on lmc membership for woh g64 ( see [ sec : mem ] ) . those exposures were 60 sec long , and obtained at the parallactic angle . each observation was also followed by a henear comparison arc . throughout the course of this and several other nights we also observed spectrophotometric standard stars from the list of hamuy et al . the spectrograph slit was oriented to the parallactic angle for these observations as well . some of those data were obtained under non - photometric conditions , but all of the standards data were combined , with a grey shift calculated such that the average through - put agreed with the observation with the highest through - put . the resulting residuals from a high - order fit were a few hundrenths of a magnitude , as expected . the sensitivity curve we derived was applied to the woh g64 observation . all data were processed using iraf . the spectra were extracted using an optimal extraction algorithm , with deviant pixels identified and rejected based upon the assumption of a smoothly varying spatial profile . before investigating the physical properties of woh g64 , we first wished to confirm that it was indeed a rsg member of the magellanic clouds , rather than an unusual foreground giant . elias et al . ( 1986 ) measure a radial velocity of 315 km s@xmath4 from several nebular emission lines , but state that the velocity zero point is uncertain by 50 km s@xmath4 , and cite this as evidence of membership in the lmc . however , this really only demonstrates that the star is not a foreground dwarf ; a halo giant will have a similar velocity to the lmc s systematic velocity , as it is due mainly to the reflex motion of the sun . this value can also be compared to the @xmath12 270 - 278 km s@xmath4 velocity of the maser emission reported by marshall et al . ( 2004 ) , which is more in line with our expectations of the lmc s rotation ( olsen & massey 2007 ) . finally , van loon et al . ( 1998 ) use the ca ii triplet to measure a heliocentric velocity of @xmath12 300 km s@xmath4 . we can use our spectra to determine if the star s radial velocity is consistent with the kinematics of other rsgs in the lmc . we measured the radial velocity of woh g64 by cross correlating our spectra against those of three lmc rsgs with accurate velocities measured by massey & olsen ( 2003 ) , [ m2002 ] lmc 022204 , 024014 , and 024987 . for this , we used the ` fxcor ` task in the iraf ` rv ` package , restricting the cross - correlation region to 8450 - 8700a , which contains the very strong triplet lines at @xmath13 ( figure 1 ) . the spectra were first normalized to unity , and then a value of 1 subtracted so that only spectral features add signal to the cross - correlation . we used the three individual woh g64 spectra in order to determine the internal consistency of our measurements ; in addition we cross - correlated each of the radial velocity standards " against each other . these tests revealed that the maximum difference seen in cross - correlating any two calibrating stars against themselves was 10 km s@xmath4 , giving us assurance that the velocity we determine with the combination of the three would be a few km / sec . we expect to be able to centroid each line to about 10% of the spectral resolution element or better , and so this is consistent since the 4.3 resolution translates to about 150 km s@xmath4 . we obtained a radial velocity of 294 @xmath0 2 km s@xmath4 , where the error refers to the formal standard deviation of the mean , and reflects the good agreement in cross - correlating the stars used for the velocity determinations against themselves . given this star s location in the lmc , we expect a heliocentric velocity of 263 km s@xmath4 due to contributions from the lmc s space motion relative to the sun and its internal rotation , using a kinematic model fit to lmc carbon stars ( olsen & massey 2007 ) . although our measured velocity for woh g64 is @xmath1230 km s@xmath4 higher than that of the carbon stars , it is still entirely consistent with lmc membership , as seen in figure [ fig : kinematics ] . the difference in velocity can be attributed to the fact that the lmc s rsgs appear to exhibit @xmath1245 km s@xmath4 faster rotation than the carbon stars ( whether this is because either the lmc s rsgs or its carbon stars have been kinematically disturbed by external forces is not yet clear ; see olsen & massey 2007 ) . in any case , woh g64 s velocity is indeed typical for an lmc rsg at its position . elias et al . ( 1986 ) obtain a 6200 - 9200 spectrum of woh g64 from the ctio blanco 4-meter and rc spectrograph . based on this spectrum they assign the star a type of @xmath12 m7.5 , and derive an m@xmath14 of @xmath15 ( adopting a distance modulus of 18.6 ) by integrating over ir ground - based and iras data and assuming the ( @xmath1 ) color to be @xmath12 8 ; we correct this to m@xmath16 ( @xmath17 ) adopting a distance modulus of 18.5 ( van den bergh 2000 ) . this work is also the first to note the incredible excess ir emission from woh g64 and postulate that such emission is coming from a substantial circumstellar dust shell . van loon et al . ( 2005 ) examine a low - resolution long - slit optical spectrum obtained with dfosc at the 1.5 m danish telescope at la silla in december 1995 . they assign a spectral type of m7.5e i , in agreement with elias et al . ( 1986 ) , but also find some disagreement between different tio bands , with the 6200 band suggesting an m7 - 8 i type while some relative band strengths suggest an earlier type of m5 i. they assign a t@xmath6 of 3008 k based on a giant - class t@xmath6 scale published in fluks et al . ( 1994 ) and use of the dust radiative transfer model dusty ( ivezi , nenkova , & elitzur 1999 ) . finally , they determine m@xmath18 ( @xmath19 ) , adopting a distance to the lmc of 50 kpc . most recently , ohnaka et al . ( 2008 ) use 2005 and 2007 n - band observations to compute 2d models of the dusty torus surrounding woh g64 . while ohnaka et al . ( 2008 ) begin by discussing the parameters derived by van loon et al . ( 2005 ) and elias et al . ( 1986 ) , the temperatures cited within are not derived from the latter papers ; instead , ohnaka et al . ( 2008 ) correlates the spectral type range of m5 - 7 with t@xmath6 values of 3200 - 3400 k. adopting the 3200 k t@xmath6 in their radiative transfer modeling , and a distance to the lmc of 50 kpc , they determine an m@xmath14 of @xmath20 ( @xmath21 ) based on radiative transfer modeling of the dusty torus around the star . they also compute 2d models with t@xmath6 = 3400 k and find that this affects the resulting parameters very little ( ohnaka et al . the final luminosity determined from this modeling is a factor of 2 lower than elias et al . ( 1986 ) and van loon et al . ( 2005 ) s luminosities . based on the strengths of the tio bands in our woh g64 optical spectrum , we assign a spectral type of m5 i by comparison with other rsgs we have classified ( papers i and ii ; massey et al . this type confirms woh g64 as the latest - type rsg in the lmc , although it is considerably earlier than the m7.5 types discussed in elias et al . ( 1986 ) and van loon et al . ( 2005 ) . the observed sed for woh g64 was compared to marcs stellar atmosphere models of metallicity @xmath22 , corresponding to the metallicity of the lmc ( westerlund et al . 1997 , massey et al . the marcs model seds available ranged from 3000 to 4500 k in 100 k increments , and were interpolated for intermediate temperatures at 25 k increments . the log @xmath23 values ranged from @xmath24 to @xmath25 in increments of 0.5 dex . when fitting the models to the data , they were reddened adopting the cardelli et al . ( 1989 ) @xmath26 reddening law . the reddening and t@xmath6 were determined by finding the best by - eye fit between the marcs models and the woh g64 sed . the reddening was based on the agreement between the model and observed continuum , while t@xmath6 was based on the strengths of the tio bands at @xmath276158 , 6658 , 7054 ( jaschek & jaschek 1990 ) , with the bluer tio bands at @xmath275167 , 5448 , 5847 used as secondary confirmations of the fit quality ; this ensured that there was minimal degeneracy in determining the best model fit . the @xmath288433 and @xmath288859 tio bands are present in the spectrum as well ; however , we have found in past work that these features are not generally well - matched by the marcs models ( see papers i and ii ) . considering woh g64 s strong tio band strengths and corresponding late spectral type , our precision for this fit was @xmath0 25 k. the extinction value @xmath29 is determined to approximately 0.15 mag precision . from fitting the 5000 - 9000 optical spectrum , we find a t@xmath6 of 3400 @xmath0 25 k and @xmath29 = 6.82 @xmath0 0.15 mag , adopting a log @xmath23 value of @xmath30 dex . our best model fit is shown in figure [ fig : spec ] , and alternative model fits demonstrating the precision of these derived properties are shown in figure [ fig : spec2 ] . in order to confirm that this is the appropriate log @xmath23 value for woh g64 , we must first determine the star s radius , which in turn requires calculation of the bolometric luminosity . while rsgs in the lmc are known to be variable by up to a magnitude or more in @xmath7 ( levesque et al . 2007 ) , their variability at @xmath31 is much smaller ( @xmath120.2 mag ; josselin et al . 2000 ) . for this reason , we determine the bolometric luminosity based on woh g64 s @xmath31 magnitude . there are a number of @xmath31 band observations for this star , including @xmath32 = 6.85 from 2mass , @xmath31 = 6.85 from buchanan et al . ( 2006 ) , @xmath31 = 6.88 from elias et al . ( 1986 ) , and @xmath31 = 6.91 from denis ( epchtein et al . the consistency of these magnitudes confirms that the @xmath31 magnitude of woh g64 has stayed quite constant with time . we adopt the 2mass @xmath32 magnitude for this work , correcting @xmath32 to @xmath31 by the relation @xmath33 ( carpenter 2001 ) and adopting an error of @xmath34 mag ( josselin et al . we adopt the t@xmath6-dependent bolometric correction at @xmath31 for the lmc ( paper ii ) : @xmath35 by taking these parameters , calculating @xmath36 ( schlegel et al . 1998 ) , and adopting a distance modulus of 18.5 ( 50 kpc ) for the lmc ( van den bergh 2000 ) with an error of @xmath37 mag , we determine a physical log @xmath23 of @xmath38 dex , which agrees with our model surface gravity . with these values , and taking @xmath39 = 4.74 ( bessell et al . 1998 ) we also determine an m@xmath40 ( log@xmath41 ) , or @xmath120.6 mag more luminous than the value obtained by ohnaka et al . we also measure a radius of r / r@xmath42 , with an uncertainty of @xmath125% . while van loon et al . ( 2005 ) mention some disparity in spectral type determinations between different tio bands , we find no such disagreement when assigning our spectral type and t@xmath6 . as a self - consistency check , we also determine these physical properties based on the @xmath43 color of woh g64 , as we have done in past papers ( papers i and ii , levesque et al . 2007 , massey et al . 2009 ) . for a @xmath7 magnitude we adopt @xmath7 = 18.63 from the macho photometric data ( allsman & axelrod 2001 ) . we deredden the photometry using @xmath44 ( schlegel et al . 1998 ) . from this we find that ( @xmath1)@xmath45 for woh g64 is 5.74 . in paper ii we determined the theoretical relation between t@xmath6 and @xmath43 based on the marcs models and the assumptions of bessell et al . ( 1998 ) concerning the effective broad - band bandpasses , and found @xmath46 with these we determined a t@xmath6 of 3372 k. once again using the bolometric correction at @xmath31 and @xmath47 and propagating our errors for these values , these numbers in turn produce a log @xmath48 dex and m@xmath49 ( log@xmath50 ) . based on these parameters we also calculate a radius of r / r@xmath51 with an uncertainly of @xmath125% . this shows excellent agreement between the parameters derived from spectral fitting as compared to the ( @xmath1)@xmath45 colors . the effective temperatures from both spectral fitting and @xmath43 colors agree with the predictions of the t@xmath6 scale for the lmc ( paper ii ) ; woh g64 is the only m5 i rsg known in the lmc , and this additional data point agrees nicely with the general trend that the lmc t@xmath6 scale is @xmath12 50 k cooler than the galactic scale ( the single m5 i rsg in the milky way , @xmath11 her , has a t@xmath6 of 3450 k ; paper i ) . the parameters determined from past work , our spectral fitting , and our ( @xmath1)@xmath45 colors are summarized in table [ tab : params ] , where we also compare our values to those derived by others . in our analysis of woh g64 , we find an unusually high bolometric luminosity , @xmath52 ( @xmath53 ) compared to the other most luminous rsgs in the magellanic clouds ( @xmath54 ) . as shown in table 1 , our value is consistent with that of elias et al . ( 1986 ) and van loon et al . ( 2005 ) , but about 0.5 mag more luminous than that found by ohnaka et al . ( 2008 ) . as is obvious from the large extinction , woh g64 is surrounded by a dense circumstellar envelope . if this envelope were spherical , then the flux absorbed along the line of sight would be re - radiated towards us in the nir , but with an asymmetrical circumstelar environment ( either a torus or a disk ) , we may receive more ( or less ) than what is absorbed . if we adopt the geometrical model proposed by ohnaka et al . ( 2008 ) , that we view a torus almost head on , then we are receiving more flux form the combined system ( star plus torus ) than is being absorbed . based on figure 5 in ohnaka et al . ( 2008 ) , we estimate the torus contributes approximately 0.5 mag at @xmath31 . in order to derive the stellar flux , we then need to correct our @xmath55 derived from @xmath31 ( -9.4 ) by this amount , and arrive at an m@xmath56 . we list our final adopted parameters in table 1 reddening law , which is certainly a possibility . however , it is also true that the argument by humphreys et al . ( 2007 ) rests on the unproven ( and unstated ) assumption of spherical symmetry for the dust emission . ] we believe our value for the extinction is to be preferred over that of the ohnaka et al . ( 2008 ) analysis , as they adopt a black - body flux distribution for the star . the use of a black - body overestimates the stellar flux in the optical , which leads to an overestimate of the amount of attenuation required to reproduce the observed data . in the end , though , our estimates for the bolometric luminosities are in excellent agreement , as shown in table 1 . ohnaka et al . ( 2008 ) did adopt a somewhat lower t@xmath6 than what we found ( 3200 k vs 3400 k ) , but fortuitously they also applied their analysis with a 3400 k t@xmath6 and derive essentially the same bolometric luminosity ( ohnaka 2009 , private communication ) . with our determination of @xmath57=3400 k and @xmath58 , we derive a stellar radius of 1540r@xmath59 . given the _ minimum _ formal errors ( @xmath60=25 k and @xmath61 mag ) , the uncertainty of the radius is 5% . this value of @xmath62 _ is _ significantly larger than the largest magellanic cloud stars ( 1240 - 1310@xmath63 ) , as expected from its cooler temperature and higher bolometric luminosity . but , while the star may be a behemoth by magellanic cloud standards , it is about the same size as the largest rsgs we found in milky way ( paper i ) , where kw sgr , case 75 , ky cyg , and @xmath64 cep all have essentially the same large radii as woh g64 . in figure [ fig : hrd ] we show the position of the star in h - r diagram , compared to the location of the geneva evolutionary tracks . we have included too the locations of the star from van loon et al . ( 2005 ) and ohnaka et al . we see that the star is cooler than the evolutionary tracks allow , but that other lmc stars also suffer from this problem ; we agree with the ohnaka et al . ( 2008 ) conclusion that this star is on the edge ( or slightly to the right of ) of the hayashi limit . the luminosity is in good agreement with the prediction of stellar evolutionary theory , and we see that the mass corresponding to this luminosity is about @xmath65 ( note that although the masses labeled are the stars _ initial _ masses , mass - loss is expected to remove only a few percent of a star s mass during its earlier evolution for stars in this mass range . ) combined with our radius estimate , this leads to a @xmath66 value of -0.5 , consistent with the surface gravity adopted in our fit ( [ sec : fit ] ) . given the 5% uncertainty in the radius , and adopting a 20% uncertainty in the mass , leads to an uncertainty of 0.1 dex in @xmath66 . elias et al . ( 1986 ) report detection of [ ] @xmath286300 , h@xmath11 , [ ] @xmath286548 , and [ ] @xmath286584 emission lines in their @xmath126 resolution spectrum of woh g64 , along with a tentative detection of [ ] @xmath276717 , 6731 emission . using the [ ] , h@xmath11 , and [ ] lines they measure a radial velocity of 315 km s@xmath4 , which implies a redshift compared to the 270 - 278 km s@xmath4 value found in the maser analysis by marshall et al . ( 2004 ) and our 294 km s@xmath4 measurement from the lines ( [ sec : mem ] ) , given that the elias et al . ( 1986 ) value may have a zero - point error as large as 50 km s@xmath4 . given that the elias et al . ( 1986 ) values have as much as 50 km / sec zero - point error , the significance of this redshift was not large , but we confirm the redshift below , using more accurate measurements . in the red part of our spectrum , we detect [ ] @xmath286300 , h@xmath11 , [ ] @xmath286548 , [ ] @xmath286584 , and [ ] @xmath286731 emission ( figure [ fig : elines]a ) . the [ ] @xmath286717 doublet line is not measurable , as its wavelength is coincident with one of the numerous absorption features in our spectrum . in addition , we detect h@xmath67 , [ ] @xmath275198 , 5200 , and [ ] @xmath285007 emission in the bluer region of the spectrum ( figure [ fig : elines]b ) . we also note that some of the tio band heads are in weak emission . this is also seen in at least one other rsg with an extreme circumstellar environment , e.g. , vy cma ( hyland et al . 1969 ; wallerstein & gonzalez 2001 ) . the lines are clearly spatially coincident with the stellar spectrum , and there is no evidence of extended ( nebular ) emission in our good seeing spectra . we would readily detect an extension of 0.3 pixels , corresponding to 0.33 " , or 0.08 pc at the distance of the lmc . we have measured the radial velocities of the emission lines by fitting gaussians to them , and find an average ( heliocentric ) value of @xmath68 km s@xmath4 , where the error refers to the standard deviation of the mean ; radial velocities for each of these features are given in table [ tab : elines ] . this value is significantly larger than the @xmath69 km s@xmath4 we measure for the star itself ( [ sec : mem ] ) . it is possible that the redshift of these emission lines can be explained in part by scattering of the spectrum by an extended , expanding dust shell , a phenomenon that has previously been observed and described in red supergiants ( romanik & leung 1981 , and references therein ) . we measure the integrated fluxes of these emission lines , and correct these values for extinction based on the h@xmath11/h@xmath67 emission line ratio , assuming the balmer decrement for case b recombination ( h@xmath11/h@xmath67 = 2.87 , following osterbrock 1989 ) and the cardelli et al . ( 1989 ) reddening law . we calculate an @xmath29 = 1.90 , very different than the value determined for the star from our spectral fitting ; however , it is not surprising that the gas we observe has a net extinction that is lower than the star , and this is in fact consistent with it forming in the outer part of the dust shell . the observed and dereddened integrated fluxes are included in table [ tab : elines ] . with these dereddened fluxes we are able to calculate several common emission line diagnostic ratios , notable log ( [ ] @xmath286584/h@xmath11 ) = 0.05 , log ( [ ] @xmath286300/h@xmath11 ) = @xmath70 , and log ( [ ] @xmath285007/h@xmath67 ) = @xmath71 ) , and a lower limit of log([]/h@xmath11 ) @xmath72 ( see baldwin et al . 1981 for a detailed discussion of these diagnostics ) . the value of the [ ] /h@xmath11 ratio is much higher than the minimum value at which [ ] is considered present " , defined in baldwin et al . ( 1981 ) as log ( [ ] @xmath286300/h@xmath11 ) = @xmath73 . this high relative strength of the [ ] line implies that the dominant source of ionization is likely shock heating . this is also implied by the relatively high lower limit of the log([]/h@xmath11 ) @xmath72 ratio in our spectrum ( see , for example , allen et al . 1999 , 2008 ) , and we thank the anonymous referee for pointing this out . however , this is not the only explanation , as called to our attention in thoughtful comments by nathan smith ( 2009 , private communication ) and phil bennett ( 2009 , private communication ) . bennett argues in particular that such nebular emission is typical of vv cep - like systems , i.e. , formed by ionization by a hot companion , but see discussion below . we also find that nitrogen is highly enhanced in the nebula ; if we use the standard galaxy diagnostics of pettini & pagel ( 2004 ) , which rely upon a constant n / o ratio , we would derive a 12 + log(o / h ) = 8.9 , considerably in excess of the 12 + log(o / h ) = 8.4 that is typical of hii regions in the lmc ( russell & dopita 1990 ) . but , far more likely is the possibility that n is simply enhanced in this gas . without other line diagnostics ( such as [ ] @xmath283727 , [ ] @xmath284363 , or [ ] @xmath286717 ) we lack a quantitative knowledge of the oxygen abundance , electron density , and electron temperature . such n enrichment is a signature of many shells and rings around massive stars , such as sher 25 ( brandner et al . 1997 , hendry et al . 2008 ) . sher 25 is a blue supergiant , but is thought to be in a blue loop phase , with the high n - enriched gas ejected during a prior evolutionary phase ( brandner et al . 1997 ; see also discussion in smartt et al . 2002 and hendry et al . other such examples are also known ( hunter et al . 2008 ) . we note that [ ] @xmath74 is strongly present . like the [ ] @xmath75 line , this doublet is also collisionally excited ( osterbrock 1989 ) , and we take its strength as further evidence of shocks and nitrogen enrichment , consistent with the above . we have used the marcs stellar atmospheres and moderate resolution spectrophotometry and broad - band photometry to investigate the physical properties of the enigmatic supergiant woh g64 . in our addition our study conclusively shows that the star is a supergiant member of the lmc , rather than ( say ) an unusual giant star in our own galaxy s halo . our analysis finds an effective temperature of 3400 @xmath0 25 k , considerably warmer than previous studies have adopted . with our parameters from spectral fitting , and taking into account the additional flux contribution at @xmath31 from the asymmetric circumstellar dust envelope as seen in ohnaka et al . ( 2008 ) , we determine m@xmath76 for woh g64 . while woh g64 has previously been cited as having a t@xmath6 and m@xmath14 that would make it the largest rsg known , those parameters have also placed it at odds with the position of the evolutionary tracks on the h - r diagram . by contrast , our parameters place woh g64 closer to the predictions of stellar evolutionary theory and in good agreement with other rsgs in the lmc . we determine a stellar radius of @xmath3 = 1540 which , while still the largest in the lmc , is now in good agreement with the size of other large rsgs ( @xmath77r@xmath5 ; paper i ) , given that such radii have 5% errors or larger . the question still remains as to whether or not woh g64 is a levesque - massey variable . its extreme physical properties and position on the h - r diagram are similar to those observed in levesque - massey variables . however , there is currently no evidence for spectral variability based on past observations ; spectra obtained by elias et al . ( 1986 ) , van loon et al . ( 2005 ) , and this work all show woh g64 remaining in a consistent , cool , late - m - type state . the disagreement in type between m5 and m7 - 8 is almost certainly a result of differing methods rather than changes in the spectrum . however , more frequent observations of woh g64 could well uncover variations in its @xmath7 magnitude and t@xmath6 ; data from the all sky automated survey ( asas ) project ( pojmanski 2002 ) demonstrates that the star s @xmath78 magnitude shows considerable variability , but there is no @xmath7 or @xmath31 band data available . on the other hand , the star s very high reddening , well - defined dusty torus , and considerable maser activity would be unique among the levesque - massey variables ; these properties suggest woh g64 s similarity to the milky way s vy cma , as well as other dust - enshrouded rsgs with maser activity such as nml cyg , vx sgr , and s per ( schuster et al . 2006 ) . finally , we detect a number of emission lines in the woh g64 optical spectrum , including h@xmath11 and h@xmath67 , [ ] , [ ] , [ ] , [ ] , and [ ] . the radial velocity of these features , 344 @xmath0 9 km s@xmath4 , is considerably greater than the star s radial velocity , 294 @xmath0 2 km s@xmath4 , which may be due to scattering by the expanding dust region . we also note the presence of some tio band heads in emission . this has been previously seen in the spectrum of vy cma , which is due to the extreme circumstellar environment . inspection of our data shows that the emission is not spatially extended , and the strength of the [ ] feature suggests that the dominant source of ionization for these features is shock heating . we can not reject the possibility that woh g64 is a binary system , with the nebular component contributed by a hot companion , but we do not favor this interpretation . first is the issue of the nebular radial velocities . we find a significant ( 50 km s@xmath4 ) velocity difference between the nebular emission and the stellar component , and , at first blush , this might be taken as an indication of binarity ; i.e. , with the nebula emission ( somehow ) tracking the motion of the hot component . however , a similar velocity was also noted by elias et al . while this could be a coincidence , it certainly does nt argue in favor of a binary hyptothesis . besides , in a short - period system we would not expect the ionized gas to be attached " to the hot star . rather , we prefer the explanation that the nebular redshift is the result of scattering within the expanding dust shell . we do note that if the putative hot companion shared the extinction measured from the nebular lines , rather than that of the rsg , its flux would dominate even in the red part of the spectrum , which is clearly not the case . if instead the companion shared the same high extinction as the rsg , but its hii region did not , we would still expect that the companion would dominate the flux in the blue part of the spectrum . a better spectrum going further into the blue should be able to substantiate or refute the possibility that a hot companion is present with the same extinction , as a late o - type dwarf would have a bolometric luminosity of @xmath79 ( @xmath80 ) . accounting for the bolometric corrections involved would lead to @xmath81 for the companion , and @xmath82 for the rsg . thus , at @xmath83 , the two stars would be equally bright , as @xmath84 for a late - type o star , while @xmath85 for a rsg ( see massey 1998a , 1998b , and paper i ) . however , further investigation into the source of these emission lines will likely require further spectral analyses , as well as observations of woh g64 in the bluer regions of the spectrum . we gratefully acknowledge lco s hospitality and assistance provided during our observations for this paper . this paper made use of data from the two micron all sky survey ( 2mass ) , which is a joint project of the university of massachusetts and the infrared processes and analysis center , california institute of technology , funded by the national aeronautics and space administration and the national science foundation . we thank phil bennett , george herbig , deirdre hunter , eric josselin , scott kenyon john monnier , keiichi ohnaka , nathan smith , jacco van loon , george wallerstein , and the anonymous referee for helpful comments and discussion regarding this manuscript . e. l.s participation was made possible in part by a ford foundation predoctoral fellowship . this work was supported by the national science foundation through ast-0604569 to p. m. allen , m. g. , dopita , m. a. , tsvetanov , z. i. , & sutherland , r. s. 1999 , apj , 511 , 686 allen , m. g. , groves , b. a. , dopita , m. a. , sutherland , r. s. , & kewley , l. j. 2008 , apjs , 178 , 20 allsman , r. a. & axelrod , t. s. ( for the macho collaboration ) 2001 , arxiv : astro - ph/0108444 baldwin , j. a. , phillips , m. m. , & terlevich , r. 1981 , pasp , 93 , 5 bessell , m. s. , castelli , f. , & plez , b. 1998 , a&a , 333 , 231 brandner , w. , grebel , e. k. , chu , y - h . , & weis , k. 1997 , apj , 475 , 45 buchanan , c. l. et al . 2006 , aj , 132 , 1890 cardelli , j. a. , clayton , g. c. , & mathis , j. s. 1989 , apj , 345 , 245 carpenter , j. m. 2001 , aj , 121 , 2851 choi , y. k. et al . 2008 , pasj , 60 , 1007 conti , p. s. , garmany , c. d. , & massey , p. 1986 , aj , 92 , 48 elias , j. h. , frogel , j. a. , & schwering , p. b. 1986 , apj , 302 , 679 epchtein , n. , et al . 1997 , the messenger , 87 , 27 fluks , m. a. , plez , b. , th , p. s. , et al . 1994 , a&as , 105 , 311 gustafsson , b. , bell , r. a. , eriksson , k. , & nordlund , . 1975 , a&a , 42 , 407 gustafsson , b. , edvardsson , b. , eriksson , k. , jorgensen , u. g. , nordlund , , & plez , b. 2008 , a&a , 486 , 951 gustafsson , b. , edvardsson , b. , eriksson , k. , mizumo - wiedner , m. , jorgensen , u. g. , & plez , b. 2003 , in asp conf . 288 , stellar atmosphere modeling , ed . i. hubeny , d. mihalas , & k. werner ( san francisco : asp ) , 331 hamuy , m. , suntzeff , n. b. , heathcote , s. r. , walker , a. r. , gigoux , p. , phillips , m. m. 1994 , pasp , 106 , 566 hayashi , c. & hoshi , r. 1961 , pasj , 13 , 552 hendry , m. a. et al . 2008 , mnras , 388 , 1127 herbig , g. h. 1969 , pnas , 63 , 1045 humphreys , r. m. , davidson , k. , ruch , g. , & wallerstein , g. 2005 , aj , 129 , 492 humphreys , r. m. , helton , a. l. , & jones , t. j. 2007 , aj , 133 , 2716 hunter , i. et al . 2008 , apj , 676 , l29 hyland , a. r. , becklin , e. e. , neugebauer , g. , & wallerstein , g. 1969 , apj , 158 , 619 ivezi , z. , nenkova , m. , elitzur , m. 1999 , user manual for dusty . university of kentucky internal report jaschek , c. & jaschek , m. 1990 , the classification of stars ( cambridge : cambridge univ . press ) jones , t. j. et al . 1993 , apj , 411 , 323 josselin , e. , blommaert , j. a. d. l. , groenewegen , m. a. t. , omont , a. , & li , f. l. 2000 , a&a , 357 , 225 le sidaner , p. , & le bertre , t. 1996 , a&a , 314 , 896 levesque , e. m. , massey , p. , olsen , k.a.g . , plez , b. , josselin , e. , maeder , a. , meynet , g 2005 , apj 628 , 973 levesque , e. m. , massey , p. , olsen , k.a.g . , plez , b. , josselin , e. , maeder , a. , meynet , g. 2006 , apj , 645 , 1102 levesque , e. m. , massey , p. , olsen , k. a. g. , & plez , b. 2007 , apj , 667 , 202 lucke , p. b. & hodge , p. w. 1970 , aj , 75 , 171 marshall , j. r. , van loon , j. th . , matsuura , m. , et al . 2004 , mnras , 355 , 1348 massey , p. 1998a , in the stellar initial mass function , 38th herstmonceux conference , ed . g. gilmore & d. howell ( san francisco : asp ) , 17 massey , p. 1998b , in stellar astrophysics for the local group , ed . a. aparicio , a. herrero , & f. sanchez ( cambridge : cambridge university press ) , 95 massey , p. 2002 , apjs , 141 , 81 massey , p. , bresolin , f. , kudritzki , r. p. , puls , j. , & pauldrack , a. w. a. 2004 , apj , 608 , 1001 massey , p. , levesque , e. m. , olsen , k. a. g. , plez , b. , & skiff , b. a. 2007 , apj , 660 , 301 massey , p. , levesque , e. m. , plez , b. , & olsen , k. a. g. 2008 , in iau symp . 250 , massive stars as cosmic engines , ed . f. bresolin et al . ( cambridge : cambridge univ . press ) , 97 massey , p. , levesque , e. m. , & plez , b. 2006 , apj , 646 , 1203 massey , p. , silva , d. r. , levesque , e. m. , plez , b. , & clayton , g. c. , meynet , g. , & maeder , a. 2009 , apj , submitted massey , p. , plez , b. , levesque , e. m. , olsen , k. a. g. , clayton , g. c. , & josselin , e. 2005 , apj , 634 , 1286 meynet , g. & maeder , a. 2005 , a&a , 429 , 581 monnier , j. d. , tuthill , p. g. , lopez , b. , cruzalebes , p. , danchi , w. c. , & haniff , c. a. 1999 , apj , 512 , 351 ohnaka , k. , driebe , t. , hoffman , k .- h . weigelt , g. , & wittkowski , m. 2008 , a&a , 484 , 371 olsen , k. a. g. & massey , p. 2007 , apj , 656 , l61 osterbrock , d. e. 1989 , astrophysics of gaseous nebulae and active galactic nuclei ( mill valley : university science books ) pettini , m. & pagel , b. e. j. 2004 , mnras , 348 , l59 plez , b. 2003 , in asp conf . 298 , gaia spectroscopy : science and technology , ed . u. munari ( san francisco : asp ) , 189 plez , b. , brett , j. m. , & nordlund , . 1992 , a&a , 256 , 551 pojmanski , g. 2002 , acta astron . , 52 , 397 roche , p. f. , aitken , d. k. , & smith , c. h. 1993 , mnras , 262 , 301 romanik , c. j. & leung , c. m. 1981 , apj , 246 , 935 russell , s. c. & dopita , m. a. 1990 , apjs , 84 , 93 schaerer , d. , meynet , g. , maeder , a. , & schaller , g. 1993 , a&as , 98 , 523 schlegel , d. j. , finkbeiner , d. p. , & davis , m. 1998 , apj , 500 , 525 schuster , m. t. , humphreys , r. m. , & marengo , m. 2006 , aj , 131 , 603 smartt , s. j. , lennon , d. j. , kudritzki , r. p. , rosales , f. , ryans , r. s. i. , & wright , n. 2002 , a&a , 391 , 979 smith , n. , humphreys , r. m. , davidson , k. , gehrz , r > d. , schuster , m. t. , & krautter , j. 2001 , aj , 121 , 1111 smith , n. 2004 , mnras , 349 , l31 van den bergh , s. 1983 , apj , 183 , l123 van loon , j. th . , zijlstra , a. a. , bujarrabal , v. , & nyman , l .- . 1996 , a&a , 306 , l29 van loon , j. th . , te lintel hekkert , p. , bujarrabal , v. , zijlstra , a. a. , & nyman , l .- . 1998 , a&a , 337 , 141 van loon , j. th . , zijlstra , a. a. , bujarrabal , v. , nyman , l .- . 2001 , a&a , 368 , 950 van loon , j. th . , cioni , m .- r . l. , zijlstra , a. a. , & loup , c. 2005 , a&a , 438 , 273 wallerstein , g. , & gonzalez , g. 2001 , pasp , 113 , 954 westerlund , b. e. 1997 , the magellanic clouds ( cambridge : cambridge univ . press ) westerlung , b. e. , olander , n. , & hedin , b. 1981 , a&as , 43 , 267 wood , p. r. , bessell , m. s. , & whiteoak , j. b. 1986 , apj , 306 , l81 l c c c c c c elias et al . 1986 & m7.5 & & -9.6 ( 5.75)&&@xmath123 & + van loon et al . 2005 & m5 , m7 - 8 & 3008 & -9.5 ( 5.70)&2575 & & + ohnaka et al . 2008 & & 3200 & -8.9 ( 5.45)&1730 & @xmath1210 & + ohnaka et al . 2008 & & 3400 & -8.9 ( 5.45)&1730 & ? & + this work , ( @xmath1)@xmath45 colors & m5 & 3372 & -9.4 ( 5.65)&1990 & 6.82 & -0.7 + this work , spectral fitting & m5 & 3400 & -9.4 ( 5.65)&1970 & 6.82 & -0.7 + this work , final preferred parameters & m5 & 3400 & -8.9 ( 5.45)&1540 & 6.82 & -0.5 + l c c c h@xmath67 & 330 & 0.82 & 6.5 + @xmath86}$ ] 5007 & 306 & 0.24 & 1.7 + @xmath87}$ ] 5199/5202 & & 2.12 & 13.9 + @xmath88}$ ] 6300 & 336 & 1.53 & 6.7 + @xmath89}$ ] 6548 & 371 & 0.72 & 3.0 + h@xmath11 & 355 & 4.31 & 17.8 + @xmath89}$ ] 6584 & 385 & 4.86 & 20.0 + @xmath90}$ ] 6731 & 326 & 1.59 & 6.3 +	woh g64 is an unusual red supergiant ( rsg ) in the large magellanic cloud ( lmc ) , with a number of properties that set it apart from the rest of the lmc rsg population , including a thick circumstellar dust torus , an unusually late spectral type , maser activity , and nebular emission lines . its reported physical properties are also extreme , including the largest radius for any star known and an effective temperature that is much cooler than other rsgs in the lmc , both of which are at variance with stellar evolutionary theory . we fit moderate - resolution optical spectrophotometry of woh g64 with the marcs stellar atmosphere models , determining an effective temperature of 3400 @xmath0 25 k. we obtain a similar result from the star s broadband @xmath1 colors . with this effective temperature , and taking into account the flux contribution from the aysmmetric circumstellar dust envelope , we calculate log(@xmath2 ) = 5.45 @xmath0 0.05 for woh g64 , quite similar to the luminosity reported by ohnaka and collaborators based on their radiative transfer modeling of the star s dust torus . we determine a radius of @xmath3 = 1540 , bringing the size of woh g64 and its position on the h - r diagram into agreement with the largest known galactic rsgs , although it is still extreme for the lmc . in addition , we use the triplet absorption feature to determine a radial velocity of 294 @xmath0 2 km s@xmath4 for the star ; this is the same radial velocity as the rotating gas in the lmc s disk , which confirms its membership in the lmc and precludes it from being an unusual galactic halo giant . finally , we describe the star s unusual nebula emission spectrum ; the gas is nitrogen - rich and shock - heated , and displays a radial velocity that is significantly more positive than the star itself by 50 km s@xmath4 .
cross - correlations often provide us very useful information about financial markets to figure out various non - trivial and complicated structures behind the stocks as multivariate time series @xcite . actually , the use of the cross - correlation can visualize collective behavior of stocks during the crisis . as such examples , we visualized the collective movement of the stocks by means of the so - called multi - dimensional scaling ( mds ) during the earthquake in japan on march 2011 @xcite . we have also constructed a prediction procedure for several stocks simultaneously by means of multi - layer ising model having mutual correlations through the mean - fields in each layer @xcite . usually , we need information about the trend of each stock to predict the price for , you might say , ` single trading ' @xcite . however , it sometimes requires us a lot of unlearnable ` craftsperson s techniques ' to make a profit . hence , it is reasonable for us to use the procedure without any trend - forecasting - type way in order to manage the asset with a small risk . from the view point of time - series prediction , elliot _ et.al . _ @xcite made a model for the spread and tried to estimate the state variables ( spread ) as hidden variables from observations by means of kalman filter . they also estimated the hyper - parameters appearing in the model by using em algorithm ( expectation and maximization algorithm ) which has been used in the field of computer science . as an example of constructing optimal pairs , mudchanatongsuk @xcite regarded pair prices as ornstein - uhlenbeck process , and they proposed a portfolio optimization for the pair by means of stochastic control . for the managing of assets , the so - called _ pairs trading _ @xcite has attracted trader s attention . the pairs trading is based on the assumption that the spread between highly - correlated two stocks might shrink eventually even if the two prices of the stocks temporally exhibit ` mis - pricing ' leading up to a large spread . it has been believed that the pairs trading is almost ` risk - free ' procedure , however , there are only a few extensive studies @xcite so far to examine the conjecture in terms of big - data scientific approach . of course , several purely theoretical approaches based on probabilistic theory have been reported . for instance , the so - called _ arbitrage pricing theory ( apt ) _ @xcite in the research field of econometrics has suggested that the pairs trading works effectively if the linear combination of two stocks , each of which is non - stationary time series , becomes stationary . namely , the pair of two stocks showing the properties of the so - called _ co - integration _ @xcite might be a suitable pair . however , it might cost us a large computational time to check the stationarity of the co - integration for all possible pairs in a market , whereas it might be quite relevant issue to clarify whether the pairs trading is actually safer than the conventional ` single trading ' ( see for instance @xcite ) to manage the asset , or to what extent the return from the pairs trading would be expected _ etc_. with these central issues in mind , here we construct a platform to carry out and to investigate the pairs trading which has been recognized an effective procedure for some kind of ` risk - hedge ' in asset management . we propose an effective algorithm ( procedure ) to check the amount of profit from the pair trading easily and automatically . we apply our algorithm to daily data of stocks in the first section of the tokyo stock exchange , which is now available at the yahoo!finance web site @xcite . in the algorithm , three distinct conditions , namely , starting ( @xmath2 ) , profit - taking ( @xmath3 ) and stop - loss ( @xmath4 ) conditions of transaction are automatically built - into the system by evaluating the spread ( gap ) between the prices of two stocks for a given pair . namely , we shall introduce three essential conditions to inform us when we should start the trading , when the spread between the stock prices satisfies the profit - taking conditions , _ etc_. by making use of a very simple way . numerical evaluations of the algorithm for the empirical data set are carried out for all possible pairs by changing the starting , profit - taking and stop - loss conditions in order to look for the best possible combination of the conditions . this paper is organized as follows . in the next section [ sec : sec2 ] , we introduce several descriptions for the mathematical modeling of pairs trading and set - up for the empirical data analysis by defining various variables and quantities . here we also mention that the pairs trading is described by a first - passage process @xcite , and explain the difference between our study and arbitrage pricing theory ( apt ) @xcite which have highly developed in the research field of econometrics . in section [ sec : sec3 ] , we introduce several rules of the game for the trading . we define two relevant measurements to quantify the usefulness of pairs trading , namely , winning probability and profit rate . the concrete algorithm to carry out pairs trading automatically is also given in this section explicitly . the results of empirical data analysis are reported and argued in section [ sec : sec4 ] . the last section is devoted to summary . in pairs trading , we first pick up two stocks having a large correlation in the past . a well - known historical example is the pair of _ coca cola _ and _ pepsi cola _ then , we start the action when the spread ( gap ) between the two stocks prices increases up to some amount of the level ( say , @xmath2 ) , namely , we sell one increasing stock ( say , the stock @xmath5 ) and buy another decreasing one ( say , the stock @xmath6 ) at the time @xmath7 . we might obtain the arbitrage as a profit ( gain ) @xmath8 : @xmath9 when the spread decreases to some amount of the level ( say , @xmath10 ) again due to the strong correlation between the stocks , and we buy the stock @xmath5 and sell the stock @xmath6 at time @xmath11 . we should keep in mind that we used here the stock price normalized by the value itself at @xmath12-times before ( we may say ` rate ' ) as @xmath13 where we defined @xmath14 as a price of the stock @xmath5 at time @xmath15 . it is convenient for us to use the @xmath16 ( or @xmath17 ) instead of the price @xmath14 because we should treat the pairs changing in quite different ranges of price . hence , we evaluate the spread between two stocks by means of the rate @xmath16 which denotes how much percentage of the price increases ( or decreases ) from the value itself at @xmath12-times before . in our simulation , we choose @xmath18}$ ] . by using this treatment ( [ eq : def_gamma ] ) , one can use unified thresholds @xmath19 which are independent of the range of prices for all possible pairs . obviously , more relevant quantities are now not the prices themselves but the spreads for the prices of pairs . it might be helpful for us to notice that the process of the spread defined by @xmath20 also produces a time series @xmath21 , which is described as a stochastic process . in financial markets , the spread ( in particular , the bid - ask spread ) is one of the key quantities for double - auction systems ( for instance , see @xcite ) and the spread between two stocks also plays an important role in pairs trading . especially , it should be regarded as a _ first - passage process _ ( or sometimes referred to as _ first - exit process _ ) ( see for instance @xcite for recent several applications to finance ) with absorbing boundaries @xmath19 , and the times @xmath7 and @xmath11 are regarded as first - passage times . actually , @xmath7 and @xmath11 are the times @xmath15 satisfying the following for the first time @xmath22 respectively . more explicitly , these times are given by @xmath23 the above argument was given for somewhat an ideal case , and of course , we might lose the money just as much as @xmath24 where @xmath25 is a ` termination time ' satisfying @xmath26 this means that we should decide a ` loss - cutting ' when the spread @xmath27 does not shrink to the level @xmath3 and increases beyond the threshold @xmath4 at time @xmath28 . it should bear in mind that once we start the trading , we get a gain or lose , hence the above @xmath11 and @xmath29 are unified as ` time for decision ' @xmath30 by @xmath31 where we defined a unit step function as @xmath32 namely , if a stochastic process @xmath33 firstly reaches the threshold @xmath3 , the time for decision is @xmath34 , whereas if the @xmath33 goes beyond the threshold @xmath4 before shrinking to the level @xmath3 , we have @xmath35 . we already mentioned that the pairs trading is based on the assumption that the spread between highly - correlated two stocks might shrink shortly even if the two prices of the stocks exhibit a temporal spread . taking into account the assumption , here we select the suitable pairs of stocks using the information about correlation coefficient ( the pearson estimator ) for pairs in order to quantify the correlation : @xmath36 and standard deviation ( volatility ) : @xmath37 it should be noted that we also use the definition of the logarithmic return of the rescaled price @xmath38 ( see equation ( [ eq : def_gamma ] ) ) for the duration @xmath39 in ( [ eq : pearson ] ) by @xmath40 and moving average of the observable @xmath41 over the duration @xmath12 as @xmath42 at first glance , the definition of ( [ eq : pearson ] ) for correlation coefficient might look like unusual because ( [ eq : def_delta_gamma ] ) accompanying with ( [ eq : def_gamma ] ) implies that the correlation coefficient consists of the second derivative of prices . however , as we already mentioned , the ` duration ' @xmath43}$ ] appearing in ( [ eq : def_gamma ] ) is quite longer than @xmath44}$ ] , namely , @xmath45 is satisfied . hence , the price difference in ( [ eq : def_gamma ] ) could not regarded as the same derivative as in the derivative definition of ( [ eq : def_delta_gamma ] ) . therefore , the definition of correlation coefficient ( [ eq : pearson ] ) is nothing but the conventional first derivative quantity . from the view point of these quantities @xmath46 , the possible pairs should be highly correlated and the standard deviation of each stock should take the value lying in some finite range . namely , we impose the following condition for the candidates of the pairs at time @xmath7 , that is , @xmath47 and @xmath48 . thus , the total number of pairs for us to carry out the pair trading ( from now on , we call such pairs as ` active pairs ' ) is given explicitly as @xmath49 where a factor @xmath50 means that we terminate the game if ` time of decision ' @xmath30 ( see equation ( [ eq : decision_t ] ) ) becomes longer than the whole playing time @xmath51}$ ] . therefore , the number of active pairs is dependent on the thresholds @xmath52 , and we see the details of the dependence in table [ tab : tb1 ] and table [ tab : tb2 ] under the condition @xmath53 ( see @xmath54 in the tables ) . from fig . [ fig : cor ] , we are also confirmed that the number pairs @xmath55 satisfying @xmath47 and @xmath48 is extremely smaller ( @xmath56 ) than the number of combinations for all possible @xmath57 stocks , namely , @xmath58 . it might be helpful for us to notice that the pairs trading could be regarded as a ` minimal portfolio ' and it can obtain the profit even for the case that the stock average decreases . actually , it is possible for us to construct such ` market neutral portfolio ' @xcite as follows . let us consider the return of the two stocks @xmath5 and @xmath6 which are described as @xmath59 where parameters @xmath60 denote the so - called ` market betas ' for the stocks @xmath61 , and @xmath62 stands for the return of the stock average , namely , @xmath63 and here we select them ( @xmath60 ) as positive values for simplicity . on the other hand , @xmath64 appearing in ( [ eq : betai])([eq : betaj ] ) are residual parts ( without any correlation with the stock average ) of the returns of stocks @xmath61 . then , let us assume that we take a short position ( ` selling ' in future ) of the stock @xmath6 by volume @xmath65 and a long position ( ` buying ' in future ) of the stock @xmath5 . for this action , we have the return of the portfolio as @xmath66 hence , obviously , the choice of the volume @xmath65 as @xmath67 leads to @xmath68 which is independent of the market ( the average stock @xmath62 ) . we should notice that @xmath69 is rewritten in terms of the profit as follows . @xmath70 therefore , in this sense , the profit @xmath71 is also independent of the market @xmath62 . this empirical fact might tell us the usefulness of paris trading . in the arbitrage pricing theory ( apt ) @xcite , the condition for searching suitable pairs is the linear combination of ` non - stationary ' time series @xmath16 and @xmath72 , @xmath73 becomes co - integration , namely , it becomes ` stationary ' . then , the quantity possesses the long - time equilibrium value @xmath74 and we write @xmath75 with a small deviation @xmath76 from the mean @xmath74 . therefore , we easily find @xmath77 namely , we obtain the profit with a very small risk . hence , the numerical checking for the stationarity of the linear combination @xmath73 by means of , for instance , exponentially fast decay of the auto - correlation function or various types of statistical test might be useful for us to select the possible pairs . however , it computationally cost us heavily for large - scale empirical data analysis . this is a reason why here we use the correlation coefficients and volatilities to investigate the active pairs instead of the co - integration based analysis as given in the references @xcite . in this section , we explain rules of our game ( trading ) using the data set for the past three years 2010 - 2012 including 2009 to evaluate the quantities like correlation coefficient and volatility in 2010 by choosing @xmath78 [ days ] . in following , we explain how one evaluates the performance of pairs trading according to the rules . obviously , the ability of the asset management by pairs treading is dependent on the choice of the thresholds @xmath19 . hence , we should investigate how much percentage of total active pairs can obtain a profit for a given set of @xmath19 . to carry out the empirical analysis , we define the ratio between the profit @xmath79 and the loss @xmath80 for the marginal spread , namely , @xmath81 as @xmath82 where @xmath83 is a control parameter . it should be noted that for positive constants @xmath84 the gap of the spreads ( profit ) @xmath85 is written as @xmath86 therefore , the difference @xmath87 appearing in the denominator of equation ( [ eq : def_alpha ] ) gives a lower bound of the profit . although the numerator @xmath88 in ( [ eq : def_alpha ] ) has no such an explicit meaning , however , implicitly it might be regarded as a ` typical loss ' because the actually realized loss @xmath89 fluctuates around the typical value and it is more likely to take a value which is close to @xmath88 . hence , for @xmath90 , the loss for the marginal spread @xmath88 is larger than the lowest possible profit once a transaction is taken place and vice versa for @xmath91 . if we set @xmath91 , it is more more likely to lose the money less than the lowest bound of the profit @xmath87 , however , at the same time , it means that we easily lose due to the small gap between @xmath4 and @xmath2 . in other words , we might frequently lose with a small amount of losses . on the other hand , if we set @xmath90 , we might hardly lose , however , once we lose , the total amount of the losses is quite large . basically , it lies with traders to decide which to choose @xmath90 or @xmath91 , however , here we set the marginal @xmath92 as a ` neutral strategy ' , that is @xmath93 thus , we have now only two thresholds @xmath94 for our pairs trading , and the @xmath4 should be determined as a ` slave variable ' from equation ( [ eq : balance2 ] ) . actually this constraint ( [ eq : balance2 ] ) can reduce our computational time to a numerically tractable revel . under the condition ( [ eq : balance2 ] ) , we sweep the thresholds @xmath95 as @xmath96 , @xmath97 ( @xmath98 ) and @xmath99 , @xmath100 ( @xmath101 ) in our numerical calculations ( see table [ tab : tb1 ] and table [ tab : tb2 ] ) . in order to investigate the performance of pairs trading quantitatively , we should observe several relevant performance measurements . as such observables , here we define the following wining probability as a function of the thresholds : @xmath102 where we defined @xmath103 where @xmath104 are numbers of wins and loses , respectively , and the conservation of the number of total active pairs @xmath105 should hold ( see the definition of @xmath106 in ( [ eq : def_n ] ) under the condition @xmath53 ) . the bracket @xmath107 appearing in ( [ eq : pw ] ) is defined by @xmath108 we also define the profit rate : @xmath109 which is a slightly different measurement from the winning probability @xmath110 . we should notice that we now consider the case with the constraint ( [ eq : balance2 ] ) and in this sense , the explicit dependences of @xmath110 and @xmath111 on @xmath4 are omitted in the above descriptions . we also keep in mind that @xmath112 takes a positive value if we make up accounts for taking the arbitrage at @xmath34 . on the other hand , the @xmath112 becomes negative if we terminate the trading due to loss - cutting . therefore , the above @xmath111 denotes a total profit for a given set of the thresholds @xmath113 . we shall list the concrete algorithm for our empirical study on the pairs trading as follows . 1 . we collect a pair of stocks @xmath114 from daily data for the past one year . 2 . do the following procedures from @xmath115 to + @xmath116 . 1 . calculate @xmath117 and @xmath118 to determine whether the pair @xmath114 satisfy the start condition . + * start condition : * * if @xmath119 and @xmath120 and @xmath121 , go to ( c ) . * if not , go to ( b ) . 2 . @xmath122 and back to ( a ) . @xmath122 and go to the termination condition . + * termination condition : * * if @xmath123 ( we ` win ' ) , go to the next pairs @xmath124 . * if not , go back to ( c ) . if @xmath125 , we ` lose ' . if @xmath126 , go to 1 then , we repeat the above procedure for all possible pairs of @xmath0 stocks listed on the first section of the tokyo stock exchange leading up to totally @xmath127 pairs . we play our game according to the above algorithm for each pair , and if a pair @xmath114 passes their decision time @xmath30 resulting in the profit : @xmath128 or the loss : @xmath129 we discard the pair @xmath114 and never ` recycle ' the pair again for pairs trading . of course , such treatment might be hardly accepted in realistic pairs trading because traders tend to use the same pairs as the one which gave them a profit in the past markets . nevertheless , here we shall utilize this somewhat ` artificial ' treatment in order to quantify the performance of paris trading through the measurements @xmath110 and @xmath111 systematically . we also simplify the game by restricting ourselves to the case in which each trader always makes a trade by a unit volume . in the next section , we show several result of empirical data analysis . here we show several empirical data analyses done for all possible pairs of @xmath0 stocks listed on the first section of the tokyo stock exchange leading up to @xmath127 pairs . the daily data sets are collected for the past four years 2009 - 2010 from the web cite @xcite . in our empirical analysis , we set @xmath78 [ days ] , @xmath130 . before we show our main result , we provide the two empirical distributions for the correlation coefficients and volatilities , which might posses very useful information about selecting the active pairs . we also discuss the distribution of the first - passage time to quantify the processing time roughly . in fig . [ fig : cor ] , we plot the distributions of @xmath131 ( left ) and @xmath132 ( right ) for the past four years ( 2009 - 2012 ) . ( left ) and volatility @xmath132 ( right ) for the past four years data sets : 2009 , 2010 , 2011 and 2012 . on the other hand , the distribution of the volatility is almost independent of year and it possess a single peak around @xmath133 . , title="fig:",width=219 ] ( left ) and volatility @xmath132 ( right ) for the past four years data sets : 2009 , 2010 , 2011 and 2012 . on the other hand , the distribution of the volatility is almost independent of year and it possess a single peak around @xmath133 . , title="fig:",width=219 ] from the left panel , we find that the distribution of correlation coefficients is apparently skewed for all years and the degree of skewness in 2011 is the highest among the four due to the great east japan earthquake as reported in @xcite . actually , we might observe that most of stocks in the multidimensional scaling plane shrink to a finite restricted region due to the strong correlations . on the other hand , the distribution of the volatility is almost independent of the year and possess a peak around @xmath133 . we are confirmed from these empirical distributions that the choice of the system parameters @xmath130 could be justified properly in the sense that the number of pairs @xmath114 satisfying the criteria @xmath134 and @xmath135 is not a vanishingly small fraction but reasonable number of pairs ( @xmath136 ) can remain in the system . we next show the distributions of the first - passage times for the data set in 2010 . it should be noted that we observe the duration @xmath15 as a first passage time from the point @xmath137 in time axis , hence , the distributions of the duration @xmath15 are given for @xmath138 respectively . we plot the results in fig . [ fig : fg0 ] . of first - passage times the left panel is @xmath140 , whereas the right is @xmath141 . we find that one confirms the lose by loss - cutting by 50 days after the start point @xmath137 in most cases and the decision is disclosed at latest by 250 days after the @xmath137 . on the other hand , we win within several days after the start and a single peak is actually located in the short time frame . , title="fig:",width=219 ] of first - passage times @xmath139 . the left panel is @xmath140 , whereas the right is @xmath141 . we find that one confirms the lose by loss - cutting by 50 days after the start point @xmath137 in most cases and the decision is disclosed at latest by 250 days after the @xmath137 . on the other hand , we win within several days after the start and a single peak is actually located in the short time frame . , title="fig:",width=219 ] from the left panel , we find that one confirms the lose by loss - cutting by 50 days after the start point @xmath137 in most cases , and the decision is disclosed at latest by 250 days after the @xmath137 . on the other hand , we win within several days after the start and a single peak is actually located in the short time frame . these empirical findings tell us that in most cases , the spread between highly - correlated two stocks actually shrink shortly even if the two prices of the stocks exhibit ` mis - pricing ' leading up to a large spread temporally . taking into account this fact , our findings also imply that the selection by correlation coefficients and volatilities works effectively to make the pairs trading useful . as our main results , we first show the wining probability @xmath110 as a function of thresholds @xmath113 defined by ( [ eq : pw ] ) in fig . [ fig : fg1 ] . to show it effectively , we display the results as three dimensional plots with contours . from these panels , we find that the winning probability is unfortunately less than that of the ` draw case ' @xmath142 in most choices of the thresholds @xmath113 . we also find that for a given @xmath143 , the probability @xmath110 is almost a monotonically increasing function of @xmath2 in all the three years . this result is naturally accepted because the trader might take more careful actions on the starting of the pairs trading for a relatively large @xmath2 . as a function of @xmath113 . from the top most to the bottom , the results in 2012 , 2011 and 2010 are plotted . the right panels are the plots for relatively small range of thresholds @xmath113 . from these panels , we find that the winning probability is less than that of the ` draw case ' @xmath142 in most cases of the thresholds @xmath113 . see also table [ tab : tb1 ] ( 2012 ) and table [ tab : tb2 ] ( 2011 ) for the details . , title="fig:",width=219 ] as a function of @xmath113 . from the top most to the bottom , the results in 2012 , 2011 and 2010 are plotted . the right panels are the plots for relatively small range of thresholds @xmath113 . from these panels , we find that the winning probability is less than that of the ` draw case ' @xmath142 in most cases of the thresholds @xmath113 . see also table [ tab : tb1 ] ( 2012 ) and table [ tab : tb2 ] ( 2011 ) for the details . , title="fig:",width=219 ] as a function of @xmath113 . from the top most to the bottom , the results in 2012 , 2011 and 2010 are plotted . the right panels are the plots for relatively small range of thresholds @xmath113 . from these panels , we find that the winning probability is less than that of the ` draw case ' @xmath142 in most cases of the thresholds @xmath113 . see also table [ tab : tb1 ] ( 2012 ) and table [ tab : tb2 ] ( 2011 ) for the details . , title="fig:",width=219 ] as a function of @xmath113 . from the top most to the bottom , the results in 2012 , 2011 and 2010 are plotted . the right panels are the plots for relatively small range of thresholds @xmath113 . from these panels , we find that the winning probability is less than that of the ` draw case ' @xmath142 in most cases of the thresholds @xmath113 . see also table [ tab : tb1 ] ( 2012 ) and table [ tab : tb2 ] ( 2011 ) for the details . , title="fig:",width=219 ] as a function of @xmath113 . from the top most to the bottom , the results in 2012 , 2011 and 2010 are plotted . the right panels are the plots for relatively small range of thresholds @xmath113 . from these panels , we find that the winning probability is less than that of the ` draw case ' @xmath142 in most cases of the thresholds @xmath113 . see also table [ tab : tb1 ] ( 2012 ) and table [ tab : tb2 ] ( 2011 ) for the details . , title="fig:",width=219 ] as a function of @xmath113 . from the top most to the bottom , the results in 2012 , 2011 and 2010 are plotted . the right panels are the plots for relatively small range of thresholds @xmath113 . from these panels , we find that the winning probability is less than that of the ` draw case ' @xmath142 in most cases of the thresholds @xmath113 . see also table [ tab : tb1 ] ( 2012 ) and table [ tab : tb2 ] ( 2011 ) for the details . , title="fig:",width=219 ] to see the result more carefully , we write the raw data produced by our analysis in table [ tab : tb1 ] ( 2012 ) and table [ tab : tb2 ] ( 2011 ) . from these two tables , we find that relatively higher winning probabilities @xmath144 are observed , however , for those cases , the number of wins @xmath145 ( or loses @xmath146 ) is small , and it should be more careful for us to evaluate the winning possibility of pairs trading from those limited data sets . in order to consider the result obtained by our algorithm for pairs trading from a slightly different aspect , we plot the profit rate @xmath111 given by ( [ eq : profit ] ) as a function of thresholds @xmath113 in fig . [ fig : fg_gain2 ] . we clearly find that for almost all of the combinations @xmath113 , one can obtain the positive profit rate @xmath147 , which means that our algorithm actually achieves almost risk - free asset management and it might be a justification of the usefulness of pairs trading . at a glance , it seems that the result of the small winning probability @xmath110 is inconsistent with that of the positive profit rate @xmath147 . however , the result can be possible to be obtained . to see it explicitly , let us assume that the pairs @xmath114 and @xmath148 lose and the pair @xmath149 wins for a specific choice of thresholds @xmath150 . then , the wining probability is @xmath151 . however , the profits for these three pairs could satisfy the following inequality : @xmath152 from the definition of the profit rate ( [ eq : profit ] ) , we are immediately conformed as @xmath153 hence , an active pair producing a relatively large arbitrage can compensate the loss of wrong active pairs by choosing the threshold @xmath113 appropriately . it might be an ideal scenario for the pairs trading . as a function of @xmath113 . from the upper left to the bottom , the results in 2012 , 2011 and 2010 are plotted . we clearly find that for almost all of the combinations @xmath113 , one can obtain the positive profit rate @xmath147 , which means that our algorithm actually achieves almost risk - free asset management and it might be a justification of the usefulness of pairs trading . , title="fig:",width=219 ] as a function of @xmath113 . from the upper left to the bottom , the results in 2012 , 2011 and 2010 are plotted . we clearly find that for almost all of the combinations @xmath113 , one can obtain the positive profit rate @xmath147 , which means that our algorithm actually achieves almost risk - free asset management and it might be a justification of the usefulness of pairs trading . , title="fig:",width=219 ] as a function of @xmath113 . from the upper left to the bottom , the results in 2012 , 2011 and 2010 are plotted . we clearly find that for almost all of the combinations @xmath113 , one can obtain the positive profit rate @xmath147 , which means that our algorithm actually achieves almost risk - free asset management and it might be a justification of the usefulness of pairs trading . , title="fig:",width=219 ] finally , we should stress that the fact @xmath154 in most cases of thresholds @xmath113 implies that automatic pairs trading system could be constructed by applying our algorithm for all possible @xmath113 in parallel . however , it does not mean that we can always obtain positive profit ` actually ' . our original motivation in this paper is just to examine ( from the stochastic properties of spreads between two stocks ) how much percentage of highly correlated pairs is suitable for the candidate in pairs trading in a specific market , namely , tokyo stock exchange . in this sense , our result could not be used directly for practical trading . nevertheless , as one can easily point out , we may pare down the candidates by introducing the additional transaction cost , and even for such a case , the game to calculate the winning probability etc . by regarding the trading as a mixture of first - passage processes might be useful . in fig . [ fig : fg_scat ] , we plot the profit rate @xmath111 against the volatilities @xmath155 as a scattergram only for the winner pairs . against the volatilities @xmath155 as a scattergram . we set the profit - taking threshold @xmath3 as @xmath156 and vary the starting threshold @xmath2 in the range of @xmath157 . we find that there exist two distinct clusters ( components ) in the winner pairs , namely , the winner pairs giving us the profit rate typically as much as @xmath158 for the range of @xmath157 , which are almost independent of @xmath155 , and the winner pairs having the profit linearly dependent on the volatility @xmath155 . note that the vertical axis is shown as ` percentage ' . , width=302 ] in this plot , we set the profit - taking threshold @xmath3 as @xmath159 and vary the starting threshold @xmath2 in the range of @xmath160 . for each active winner pair , we observe the profit rate @xmath111 and the average volatility @xmath155 of the two stocks in each pair , and plot the set @xmath161 in the two - dimensional scattergram . from this figure , we find that there exist two distinct clusters ( components ) in the winner pairs , namely , the winner pairs giving us the profit rate typically as much as @xmath162 for the range of @xmath163 , which are almost independent of @xmath155 , and the winner pairs having the profit rate linearly dependent on the volatility @xmath155 . the former is a low - risk group , whereas the latter is a high - risk group . the density of the points for the low - risk group in fig . [ fig : fg_scat ] is much higher than that of the high - risk group . hence , we are confirmed that our selection procedure of the active pairs works effectively to manage the asset as safely as possible by reducing the risk which usually increases as the volatility grows . finally , it should be noted that as we discussed in subsection [ sec : sub_constraint ] , the value @xmath164 is a lower bound of the profit rate ( see equations ( [ eq : lower_bound ] ) and ( [ eq : eta02 ] ) ) . therefore , in the above case , the lower bound for the profit rate @xmath111 should be estimated for @xmath165 as @xmath166 the lowest value for the profit rate ( [ eq : lower_bound2 ] ) is consistent with the actually observed lowest value in the scattergram shown in fig . [ fig : fg_scat ] . finally , we shall list several examples of active pairs to win the game . of course , we can not list all of the winner pairs in this paper , hence , we here list only three pairs as examples , each of which includes _ sanyo special steel co. ltd . _ ( i d : 5481 ) and the corresponding partners are _ hitachi metals . ltd . _ ( i d : 5486 ) , _ mitsui mining @xmath167 smelting co. ltd . _ ( i d : 5706 ) and _ pacific metals co. ltd . _ ( i d : 5541 ) . namely , the following three pairs @xmath168 actually won in our empirical analysis of the game . note that each i d in the above expression corresponds to each identifier used in yahoo!finance @xcite . as we expected as an example of _ coca cola _ and _ pepsi cola _ @xcite , these are all the same type of industry ( the steel industry ) . we would like to stress that we should act with caution to trade using the above pairs because the pairs just won the game in which the pairs once got a profit in the past are never recycled in future . therefore , we need much more extensive analysis for the above pairs to use them in practice . in this paper , we proposed a very simple and effective algorithm to make the pairs trading easily and automatically . we applied our algorithm to daily data of stocks in the first section of the tokyo stock exchange . numerical evaluations of the algorithm for the empirical data set were carried out for all possible pairs by changing the starting ( @xmath2 ) , profit - taking ( @xmath3 ) and stop - loss ( @xmath4 ) conditions in order to look for the best possible combination of the conditions @xmath169 . we found that for almost all of the combinations @xmath113 under the constraint @xmath170 , one can obtain the positive profit rate @xmath147 , which means that our algorithm actually achieves almost risk - free asset management at least for the past three years ( 2010 - 2012 ) and it might be a justification of the usefulness of pairs trading . finally , we showed several examples of active pairs to win the game . as we expected before , the pairs are all the same type of industry ( for these examples , it is the steel industry ) . we should conclude that the fact @xmath154 in most cases of thresholds @xmath113 implies that automatic pairs trading system could be constructed by applying our algorithm for all possible @xmath113 in parallel way . of course , the result does not mean directly that we can always obtain positive profit in a practical pairs trading . our aim in this paper was to examine how much percentage of highly correlated pairs is suitable for the candidate in pairs trading in a specific market , namely , tokyo stock exchange . in this sense , our result could not be used directly for practical pairs trading . nevertheless , we may pare down the candidates by introducing the additional transaction cost , and even for such a case , the game to calculate the winning probability etc . by regarding the trading as a mixture of first - passage processes might be useful . we are planning to consider pairs listed in different stock markets , for instance , one is in tokyo and the other is in ny . then , of course , we should also consider the effect of the exchange rate . those analyses might be addressed as our future study . . details of the result in 2012 . see also fig . [ fig : fg1 ] ( the top most ) . we find that relatively higher winning probabilities @xmath144 are observed , however , the number of wins @xmath145 ( and lose @xmath146 ) is small . it should be noted that @xmath171 holds . [ cols=">,>,>,>,>,>,>,>,>,>,>,>",options="header " , ] this work was financially supported by grant - in - aid for scientific research ( c ) of japan society for the promotion of science no . 2533027803 and grant - in - aid for scientific research ( b ) of japan society for the promotion of science no . we were also supported by grant - in - aid for scientific research on innovative area no . one of the authors ( ji ) thanks anirban chakraborti for his useful comments on this study at the early stage . t. ibuki , s. suzuki and j. inoue , _ cluster analysis and gaussian mixture estimation of correlated time - series by means of multi - dimensional scaling _ , _ econophysics of systemic risk and network dynamics , new economic windows _ vol . 2013 , pp . 239 - 259 , springer - verlag ( italy - milan ) ( 2012 ) . t. ibuki , s. higano , s. suzuki and j. inoue , _ hierarchical information cascade : visualization and prediction of human collective behaviour at financial crisis by using stock - correlation _ , _ ase human journal _ * 1 * , issue 2 , pp.74 - 87 ( 2012 ) . t. ibuki , s. higano , s. suzuki , j. inoue and a. chakraborti , _ statistical inference of co - movements of stocks during a financialcrisis _ , _ journal of physics : conference series _ * 473 * , 012008 ( 16pages ) ( 2013 ) . m. murota and j. inoue , _ characterizing financial crisis by means of the three states random field ising model _ , _ econophysics of agent - based models , new economic windows _ vol . 2014 , pp . 83 - 98 , springer - verlag ( italy - milan ) ( 2013 ) . e.g. gatev , w.n . goetzmann and k.g . rouwenhorst , _ pairs trading : performance of a relative value arbitrage rule _ , _ the review of financial studies _ * 19 * , issue 3 , pp . 797 - 827 ( 2006 ) . ( see also _ nber working papers _ * 7032 * , _ national bureau of economic research inc . _ ( 1999 ) . ) t. ibuki and j. inoue , _ response of double - auction markets to instantaneous selling - buying signals with stochastic bid - ask spread _ , _ journal of economic interaction and coordination _ * 6 * , no.2 , pp.93 - 120 ( 2011 ) .	we carry out a large - scale empirical data analysis to examine the efficiency of the so - called pairs trading . on the basis of relevant three thresholds , namely , starting , profit - taking , and stop - loss for the ` first - passage process ' of the spread ( gap ) between two highly - correlated stocks , we construct an effective strategy to make a trade via ` active ' stock - pairs automatically . the algorithm is applied to @xmath0 stocks listed on the first section of the tokyo stock exchange leading up to totally @xmath1 pairs . we are numerically confirmed that the asset management by means of the pairs trading works effectively at least for the past three years ( 2010 - 2012 ) data sets in the sense that the profit rate becomes positive ( totally positive arbitrage ) in most cases of the possible combinations of thresholds corresponding to ` absorbing boundaries ' in the literature of first - passage processes .
the quantum nature of gravity is one of the most important subjects in theoretical physics . it is getting even more important because the higgs mass measured in lhc suggests that the standard model would be valid all the way up to the planck scale . in this talk we explain the lattice gauge theory can play crucial role in the study of the quantum gravity . superstring theory is a promising candidate of the theory of everything , which unifies the standard model of particle physics and gravity . in particular , it provides us with a consistent theory of quantum gravity . in fig . [ fig : string_scattering ] scattering of two closed strings is shown . strings propagate from left to right and sweep a two - dimensional surface called ` world - sheet . ' in the left figure , two closed strings interact and merge to a single closed string , and then it splits to two . this is the tree level diagram . in the right figure , a loop of closed strings appear as an intermediate state . each closed string loop multiplies a factor @xmath1 to the amplitude , where @xmath2 is the string coupling constant ; a diagram with @xmath3 closed string loops , which is represented by a genus-@xmath3 surface , carries a factor @xmath4 . in principle , one can take into account the quantum correction order by order by increasing the genus . but it is just a perturbative definition at best . how can one formulate string theory nonperturbatively ? in 1970s t hooft pointed out a deep relationship between the yang - mills theory with the @xmath5 gauge group and string theory , based on the @xmath0 expansion around the t hooft large-@xmath6 limit ( i.e. @xmath7 fixed ) . because the gauge field is represented by @xmath8 matrices , the propagator is represented by two lines , which represent the row and column , respectively . the interaction is described by three- and four - point vertices ( fig . [ fig : interaction_vertices ] ) . a typical example of feynman diagrams is shown in fig . [ fig : dynamical_triangulation ] . it contains the vertices , propagators and closed loops of matrix indices . because the action is written as @xmath9 , each vertex and propagator are associated with @xmath6 and @xmath10 , respectively . a closed loop carries a factor @xmath6 because a sum over the index , which runs from @xmath11 to @xmath6 , is involved . therefore , each vacuum diagram is proportional to @xmath12 , where @xmath13 . as a next step , we relate feynman diagrams to two - dimensional surfaces . intuitively , a feynman diagram with many lines looks like a dense ` fishnet ' , which eventually becomes indistinguishable from a continuum two - dimensional surface . in order to make this intuition mathematically well - defined , let us notice that a feynman diagram can be translated to a two - dimensional surface made of triangles and quadrangles , as shown in fig . [ fig : dynamical_triangulation ] . from the figure it is easy to see that @xmath14 therefore @xmath15 which is the euler number of the two - dimensional surface . note that the euler number depends only on the number of genus @xmath3 , as demonstrated in fig . [ fig : sphere ] and fig . [ fig : torus ] . therefore , it is natural to identify the string world - sheet and the coupling constant @xmath2 with the feynman diagram and @xmath0 , respectively . in this picture , the closed string corresponds to the wilson loop , because gluons sourced by the wilson loop form a cylinder - shaped fishnet , which is identified with the world - sheet swiped by the closed string ( fig . [ fig : wilsonloop ] ) . if this identification is correct , then the yang - mills theory provides us with a nonperturbative formulation of string theory , because it is defined without relying on the @xmath0 expansion . ( a usual instanton effect @xmath16 translates into @xmath17 and hence it is nonperturbative also in the context of string theory . ) this argument is rather formal and it was not clear which specific gauge theory describes string theory . more than 20 years after t hooft s proposal , maldacena conjectured that maximally supersymmetric @xmath5 yang - mills theory in @xmath18 dimensions corresponds to type ii superstring theory on the black @xmath19-brane background @xcite . here the black @xmath19-brane is a black hole - like object with @xmath19-dimensional spatial extension . the black @xmath20-brane , which we consider below , is just a black hole . therefore , if maldacena s conjecture is correct , by studying the @xmath0 correction in yang - mills theory one can learn about quantum gravity . the gauge / gravity duality conjecture motivated kaplan , katz and unsal to put super yang - mills theories in various dimensions on lattice@xcite . by now there are various lattice and non - lattice techniques with which one can simulate super yang - mills theories in various dimensions . in this talk we consider the @xmath21-dimensional theory , which is numerically most tractable . we compare the mass of the black hole and its counterpart in the gauge theory and find a good agreement at finite values of @xmath6 , which provides us with quantitative evidence supporting the maldacena conjecture at finite-@xmath6 and quantum gravity level . we briefly explain the dual gravity prediction for the gauge theory in sec . [ sec : duality ] and test it by monte carlo simulation in sec . [ sec : test ] . we discuss implications of our result and future directions in sec . [ sec : discussion ] . black @xmath19-brane consists of a large number of d@xmath19-branes @xcite , which are @xmath18-dimensional objects on which open strings can have endpoints ( fig . [ fig : dbrane_1 ] ) . the low - energy dynamics of @xmath6 d - branes are described by maximally supersymmetric @xmath5 yang - mills theory in @xmath18 dimensions @xcite , whose action is given by @xmath22 ^ 2 \right\ } + ( fermions ) , \end{aligned}\ ] ] where @xmath23 run from @xmath11 to @xmath24 and @xmath25 @xmath26 are scalar fields . the spacetime dimension is @xmath24 because open strings can move only along d - branes . intuitively , @xmath27-components of the matrices represent an open string connecting @xmath28-th and @xmath29-th d - branes . scalar fields @xmath25 describe the fluctuations of d - branes and open strings to the transverse directions . d - branes are massive objects and hence source the gravitational field , or in other words , d - branes emit close strings , the spacetime gets curved , and they turn to a black @xmath19-brane . near the horizon , the metric of the black @xmath19-brane geometry is given by @xcite @xmath30 \nonumber\\ & & \qquad + \frac{g_{ym}\sqrt{d_p n}}{u^{(7-p)/2}\left(1-\frac{u_0^{7-p}}{u^{7-p}}\right)}du^2 + g_{ym}\sqrt{d_p n}u^{(p-3)/2}d\omega_{8-p}^2 \biggl\ } , \label{near_extremal_metric } \end{aligned}\ ] ] where the yang - mills coupling @xmath31 and the size of the gauge group @xmath6 in the corresponding super yang - mills theory are used . a constant @xmath32 is the square of the string length , @xmath33 represent the @xmath18-dimensional extension of the brane , @xmath34 and @xmath35 are the radial and angular coordinate of the transverse directions , and @xmath36 is the place of the horizon . the hawking temperature is @xmath37 where @xmath38 . the string coupling constant is given by @xmath39 when @xmath40 is fixed , it behaves as @xmath41 , in the same way as in t hooft s identification . maldacena conjectured that super yang - mills describes full string dynamics around the black brane and the @xmath0 correction corresponds to the quantum correction . the hawking temperature and the mass of the black brane are identified with the temperature and the energy @xmath42 of the gauge theory . in this work we study the case of @xmath43 . the gauge theory is quantum mechanics of @xmath8 matrices , which was originally proposed as the matrix model of m - theory @xcite . the black hole mass can be calculated by adding stringy corrections to the black @xmath20-brane geometry shown above . the result is @xcite @xmath44 where @xmath45 is dimensionless effective temperature and @xmath46 are unknown constants . the energy @xmath47 is also made dimensionless , by multiplying @xmath48 . in the following , we simply set @xmath49 without loss of generality . higher order terms in each power of @xmath0 represent to the @xmath32 corrections , which appear because strings are not point - like . if maldacena s conjecture is correct , this expression must be reproduced from the matrix quantum mechanics . the @xmath50 terms of the dual gravity prediction ( [ gravity_prediction ] ) have been tested previously in @xcite . in particular , the @xmath32 correction @xmath51 has been confirmed with @xmath52 , by looking at @xmath53 @xcite . in order to study the @xmath0 correction , we must study rather small values of @xmath6 . here we study @xmath54 and @xmath55 . ( that we have to take @xmath6 small caused a technical problem which required a proper treatment ; see @xcite . ) we also have to study low temperature where the @xmath32 correction becomes small , because there are too many fitting parameters otherwise . for this reason we consider @xmath56 . in fig . [ fig : n4_law ] we plot @xmath57 . remarkably , this quantity behaves as @xmath58.therefore the first two terms of the @xmath59 expansion agree with the dual gravity prediction , and terms of order @xmath60 and higher are small compared to the @xmath58 term . as a further test , we estimated the coefficient of @xmath59 by using a fitting ansatz @xmath61 . in fig.[fig : coeff ] we show this coefficient normalized by the dual gravity prediction , @xmath62 . it is reasonably close to @xmath11 . this is very strong evidence that the gauge / gravity duality holds at quantum gravity level . we have shown the first quantitative evidence for the gauge / gravity duality at the level of finite @xmath6 ( gauge theory side ) and quantum gravity ( gravity side ) beyond kinematics , by studying the finite temperature system in which supersymmetry is broken by the thermal effect . we studied mass of the black hole in superstring theory and its counterpart in the yang - mills theory , and found good agreement . it strongly suggests that , as maldacena claimed in the first paper on the gauge / gravity duality @xcite , _ the gauge theory provides us with the nonperturbative formulation of superstring theory_. then , by studying the dynamics of super yang - mills theory on computers , we can learn about quantum gravity . furthermore , in @xcite , we found that the negative specific heat appears at very low temperature ( @xmath63 for @xmath54 and @xmath55 ) . it suggests that the quantum correction drives black hole unstable and makes it evaporate . if it can be shown more robustly , then the black hole described through the gauge theory would provides us with a concrete counter - example to hawking s information loss paradox @xcite . we expect further numerical study of the yang - mills theory will lead us to a better understanding about the information puzzle , especially how the information comes back from the black hole . this talk was based on the collaboration with y. hyakutake , g. ishiki and j. nishimura @xcite . i would like to thank them for useful comments on this proceeding . m. hanada , pos lattice * 2012 * , 019 ( 2012 ) . g. t hooft , nucl . b * 72 * , 461 ( 1974 ) . j. m. maldacena , adv . * 2 * , 231 ( 1998 ) . n. itzhaki , j. m. maldacena , j. sonnenschein and s. yankielowicz , phys . d * 58 * , 046004 ( 1998 ) . d. b. kaplan , e. katz and m. nsal , jhep * 0305 * , 037 ( 2003 ) . j. polchinski , phys . lett . * 75 * , 4724 ( 1995 ) . e. witten , nucl . b * 460 * , 335 ( 1996 ) . g. w. gibbons and k. -i . maeda , nucl . b * 298 * , 741 ( 1988 ) ; g. t. horowitz and a. strominger , nucl . b * 360 * , 197 ( 1991 ) . t. banks , w. fischler , s. h. shenker and l. susskind , phys . d * 55 * , 5112 ( 1997 ) . b. de wit , j. hoppe and h. nicolai , nucl . b * 305 * , 545 ( 1988 ) . y. hyakutake , ptep * 2014 * , no . 3 , 033b04 . k. n. anagnostopoulos , m. hanada , j. nishimura and s. takeuchi , phys . * 100 * , 021601 ( 2008 ) . s. catterall and t. wiseman , phys . d * 78 * , 041502 ( 2008 ) . m. hanada , y. hyakutake , j. nishimura and s. takeuchi , phys . lett . * 102 * , 191602 ( 2009 ) .	the strongest version of the gauge / gravity duality conjecture relates the @xmath0 correction in super yang - mills theory and the quantum correction in superstring theory . we perform a quantitative test of this conjecture at finite temperature , by studying the d0-brane matrix quantum mechanics and the black zero - brane in type iia superstring theory . we find good agreement , which strongly suggests that the super yang - mills theory does provide us with a nonperturbative formulation of _ quantum _ gravity through the gauge / gravity duality .
in this work we will determine the mosaic number of all 36 prime knots of eight crossings or fewer . before we do this , we will give a short introduction to knot mosaics . take a length of rope , tie a knot in it , glue the ends of the rope together and you have a _ mathematical knot _ a closed loop in 3-space . a rope with its ends glued together without any knots is referred to as the _ trivial knot _ , or just an unknotted circle in 3-space . there are other ways to create mathematical knots aside from rope . for example , _ stick knots _ are created by gluing sticks end to end until a knot is formed ( see adams @xcite ) . in 2008 , lomonaco and kauffman @xcite developed an additional structure for considering knots which they called _ knot mosaics_. in 2014 , kuriya and shehab @xcite showed that this representation of knots was equivalent to tame knot theory , or knots with rope , implying that tame knots can be represented equivalently with knot mosaics . this means any knot that can be made with rope can be represented equivalently with a knot mosaic . + a _ knot mosaic _ is the representation of a knot on an @xmath0 grid composed of 11 tiles as depicted in figure [ f : tiles ] . a tile is said to be _ suitably connected _ if each of its connection points touches a connection point of a contiguous tile . several examples of knot mosaics are depicted in figure [ f : example ] . it should be noted that in figure [ f : example ] , the first mosaic is a knot , the trefoil knot , the second mosaic is a link , the hopf link , and the third is the composition of two trefoil knots ( remove a small arc from two trefoils then connect the four endpoints by two new arcs depicted in red , denoted by @xmath1 ) . a knot is made of one component ( i.e. one piece of rope ) , and a link is made of one or more components ( i.e. one or more pieces of rope ) . for this work , we will focus on knot mosaics of _ prime knots_. a prime knot is a knot that can not be depicted as the composition of two non - trivial knots . the trefoil is a prime knot . @xmath2 respectively . ] + + + + when studying knots , a useful and interesting topic used to help distinguish two knots is _ knot invariants_. a knot invariant is a quantity defined for each knot that is not changed by _ ambient isotopy _ , or continuous distortion , without cutting or gluing the knot . one such knot invariant is the _ crossing number _ of a knot . the crossing number is the fewest number of crossings in any diagram of the knot . for example , the crossing number of the trefoil is three , which can be seen in figure [ f : example ] . a _ reduced _ diagram of a knot is a projection of the knot in which none of the crossings can be reduced or removed . the fourth knot mosaic depicted in figure [ f : example ] is an example of a non - reduced trefoil knot diagram . in this example the crossing number of three is not realized because there are two extra crossings that can be easily removed . + an interesting knot invariant for knot mosaics is the _ mosaic number_. the mosaic number of a knot @xmath3 is the smallest integer @xmath4 for which @xmath3 can be represented on an @xmath0 mosaic board . we will denote the mosaic number of a knot @xmath3 as @xmath5 . for the trefoil , it is an easy exercise to show that mosaic number for the trefoil is four , or @xmath6 . to see this , try making the trefoil on a @xmath7 board and arrive at a contradiction . + next , we introduce a technique that can be used to clean up " a knot mosaic by removing unneeded crossing tiles . in 1926 , kurt reidemeister demonstrated that two knot diagrams belonging to the same knot , up to ambient isotopy , can be related by a sequence of three moves , now know as the reidemeister moves @xcite . for our purposes , we will consider two of these moves on knot mosaics , the mosaic reidemeister type i and type ii moves as described by lomonaco and kauffman @xcite . for more about reidemeister moves , the interested reader should see adams @xcite . + the mosaic reidemeister type i moves are the following : the mosaic reidemeister type ii moves are given below : next we make several observations that will prove useful . [ o : twofold ] once the inner tiles of a mosaic board are suitably placed , there are only two ways to complete the board so that it is suitable connected , resulting in a knot or a link @xcite . for any given @xmath0 mosaic board , we will refer to the collection of inner tiles as the _ inner board_. for example , in figure [ f:5by5 ] the tiles @xmath8-@xmath9 would make the inner board for this mosaic board . [ o : evenall ] assume @xmath4 is even . for a board of size @xmath4 with the inner board consisting of all crossing tiles , any resulting suitably connected mosaic will either be a @xmath10 component link mosaic or @xmath11 component non - reduced link mosaic , see figure [ f:4_1 ] for example . [ o : oddall ] assume @xmath4 is odd . for a board of size @xmath4 with the inner board consisting of all crossing tiles , any resulting suitably connected mosaic is a non - reduced knot mosaic . observation [ o : oddall ] can be generalized in the following way . [ o : corner ] let @xmath12 be a knot mosaic with two corner crossing tiles in a top row of the inner board . if the top boundary of the row has an odd number of connection points , then a reidemeister type i move can be applied to either corner of the row of @xmath12 . this extends via rotation to the outer - most columns and the lower - most row of the inner board . [ fig:5by5 ] armed with this quick introduction to knot mosaics , we are ready to determine the mosaic number for all prime knots with a crossing number of eight or fewer in the next section . before we proceed , it should be noted that there are many other questions to consider regarding knot mosaics besides finding the mosaic number . for example , what is the fewest number of non - blank tiles needed to create a specific knot ? this could be known as the _ tile number _ of a knot . with this in mind , if we allow knot mosaics to be rectangular , can some knots have a smaller tile number if they are presented in a rectangular , @xmath13 configuration as opposed to a square configuration ? we will conclude this article with a number of other open questions about knot mosaics that you can consider and try to solve . + in this section , we will determine the mosaic number for all prime knots of eight or fewer crossings . we are referring to these as small " prime knots . we will see that for some knots , the mosaic number is obvious , " while others take some considerable work . we begin with knots is the second knot of six crossings in the rolfsen knot table @xcite . ] of an obvious mosaic number as shown in table [ t : obvious ] . + why do these knots have obvious mosaic numbers ? as previously noted the trefoil knot , @xmath14 , can not fit on a @xmath15 mosaic board , as such a board would only allow one crossing tile when @xmath14 requires at least three crossings . hence , the mosaic number for the trefoil is obvious . similarly , the knots @xmath16 , @xmath17 , and @xmath18 have more than four crossings , so they can not fit on a @xmath19 mosaic board which only allows at most four crossing tiles . in the appendix , we have provided representations of these knots on @xmath20 mosaic boards , thus determining the mosaic number for these knots . it should be noted that as the knots become larger , it is often difficult to determine whether a specific knot mosaic represents a given knot . to check that a knot mosaic represents a specific knot , we used a software packaged called knotscape @xcite developed by professor morwen thistlethawite which looks at the dowker notation of a knot to determine the knot presented . while knotscape can not determine all knots , it can determine small prime knots . for more see adams @xcite . + next we consider knots whose mosaic number is almost obvious . " at first glance , one may think that the figure - eight knot , @xmath21 , should have a mosaic number of four . start with a @xmath19 mosaic with the four inner tiles being crossing tiles . by observation [ o : twofold ] , this @xmath19 mosaic can be completed in two ways as seen in figure [ f:4_1 ] . however , the knot @xmath21 is known as an _ alternating _ knot . an alternating knot is a knot with a projection that has crossings that alternate between over and under as one traverses around the knot in a fixed direction . so , if we were to try to place @xmath21 on a @xmath19 mosaic , there would be four crossing tiles and they would have to alternate . thus @xmath21 can not be placed on a @xmath22 board . in the appendix we see a presentation of @xmath21 on a @xmath20 mosaic board , hence @xmath23 . another knot with an almost obvious mosaic number is @xmath24 . figure [ f : compare6_1 ] first depicts a configuration of @xmath24 on a @xmath25 mosaic board . however , by performing a move called a _ flype _ ( see adams @xcite ) we can fit @xmath24 on a @xmath20 mosaic board . it should be noted that this mosaic representation of @xmath24 has seven crossings instead of six . thus the mosaic number for @xmath24 is realized when the crossing is not . it turns out that @xmath24 is not the only knot with such a property . ludwig , evans , and paat @xcite created an infinite family of knots whose mosaic numbers were realized only when their crossing numbers were not . at this point , we have determined the mosaic number for all six or fewer crossing knots except @xmath26 . surprisingly , we will see that @xmath26 can not fit on a @xmath20 board , even though such a board has nine possible positions to place crossing tiles . + since @xmath26 has six crossings , we know that @xmath28 . in the appendix we see a representation of @xmath26 on a @xmath25 mosaic board , so @xmath29 . this mean @xmath30 or @xmath27 . we now argue @xmath27 . + assume to the contrary that @xmath31 . by the definition of the mosaic number , this implies that there is some @xmath20 mosaic @xmath12 that represents @xmath26 . we will show via a case analysis that regardless of how the crossing tiles are arranged on @xmath12 , the resulting knot is in fact _ not @xmath26_. this will give us a contradiction , implying that @xmath27 . + in order to help with the case analysis , we label the nine inner tiles of the @xmath20 mosaic @xmath32 @xmath33 , as depicted in figure [ f:5by5 ] . the @xmath26 knot uses at least six crossing tiles . since @xmath26 has a crossing number of six , by the pigeon hole principle at least one of the four corner inner tiles , @xmath32 , @xmath34 , @xmath35 , or @xmath33 must be a crossing tile . by rotations , it is enough to consider the four cases that are depicted in figure [ f : fourcases ] . note that in figure [ f : fourcases ] , gray tiles describe non - crossing tiles and white tiles could be crossing tiles ( but they do not have to be ) . * case 1 * : suppose that @xmath34 is a crossing tile , while @xmath36 , and @xmath33 are not . thus @xmath37 and @xmath38 are all crossing tiles since @xmath12 has exactly 6 crossing tiles . every reduced projection of an alternating knot is alternating , and since @xmath12 represents the alternating knot @xmath26 , the crossings on @xmath12 must alternate . a quick inspection shows that to suitably connect the inner tiles of @xmath12 we would need to ensure that ( i ) @xmath36 , and @xmath33 are not crossing tiles , ( ii ) the crossings are alternating , and ( iii ) there are no easily removed crossings , however this results in a mosaic that represents @xmath18 , as seen in figure [ fig : ex1 ] . therefore , @xmath26 can not be constructed in case 1 . * case 2 * : for the case when @xmath12 has two corner inner tiles that are crossings , we require two sub - cases . + * sub - case 2(a ) * : suppose that @xmath32 and @xmath34 are crossing tiles , while @xmath35 and @xmath33 are not . if @xmath39 is a crossing tile then observation [ o : corner ] may be applied to the top inner row of @xmath12 . applying this observaiton will either change @xmath32 or @xmath34 to a non - crossing tile . without loss of generalization , suppose that @xmath32 is changed . notice that @xmath12 now satisfies case 1 , and from the previous analysis , @xmath12 does not represent @xmath26 . therefore we may assume that @xmath39 is not a crossing tile . since @xmath12 has at least 6 crossing tiles , and @xmath40 and @xmath33 are not crossing tiles , the remaining 6 inner tiles must be crossing tiles . then @xmath39 has @xmath41 connection points . by observation [ o : corner ] either @xmath32 or @xmath34 can be changed to a non - crossing tile . again @xmath12 falls into case 1 and does not represent @xmath26 . + * sub - case 2(b ) * suppose that @xmath34 and @xmath35 are crossing tiles while @xmath32 and @xmath33 are not . + assume to the contrary that @xmath12 only has @xmath42 crossing tiles . then exactly one of @xmath43 is a non - crossing tile . up to rotation and reflection , we only need to consider situations ( i ) and ( ii ) , where since @xmath12 has 6 crossings , the corner inner - tiles @xmath34 and @xmath35 can not be changed from crossing tiles . with this in mind , suitably connecting the crossing tiles in figure [ f : case2(b)inner ] so that a knot ( and not a 2-component link ) is created leads to @xmath18 or @xmath46 ( see figure [ f : case2(b)knots ] ) . this is a contradiction , proving our claim . if the crossing tiles are alternating , then @xmath12 represents @xmath47 , contradicting that @xmath12 is @xmath26 . so assume that @xmath12 has 7 crossings and is non - alternating . + observe that if any of the pairs @xmath48 , @xmath49 , @xmath50 , or @xmath51 are non - alternating , then a type ii reidemeister move is present , and @xmath12 can be reduced to five crossings . however , this contradicts that @xmath12 is @xmath26 . so each pair @xmath48 , @xmath52 , @xmath50 , and @xmath51 is alternating . since @xmath12 is non - alternating , at least one of the pairs @xmath53 , @xmath54 , @xmath55 , or @xmath56 is non - alternating . without loss of generality , assume @xmath53 creates a pair of non - alternating crossings . then , up to ambient isotopy , @xmath12 is @xmath24 or @xmath14 , as seen in figure [ fig_6131 ] . therefore sub - case 2(b ) does not result in @xmath26 . * case 3 : * suppose that @xmath57 and @xmath33 are crossing tiles , and @xmath35 is not . since @xmath12 has at least 6 crossings , there must be at least 3 more crossing tiles on the board . + observe that if either @xmath39 or @xmath58 are crossing tiles then observation [ o : corner ] may be applied to the top inner row or the right inner column , respectively . as a result of this , one of the crossings on @xmath57 and @xmath33 can be changed to a non - crossing tile , leaving only two corner inner - tiles that are crossings . this reverts to case 2 showing that @xmath12 would not represent @xmath26 . therefore , we may assume that neither @xmath39 nor @xmath58 are crossing tiles . + with @xmath39 and @xmath58 eliminated as crossing tiles , @xmath59 and @xmath33 must all be crossing tiles . then by observation [ o : corner ] , either @xmath34 or @xmath33 can be changed from a crossing tile to a non - crossing tile . this again reverts to case 2 showing that @xmath12 would not represent @xmath26 . + * case 4 : * suppose that @xmath60 and @xmath33 are crossing tiles . note that at least one of the tiles in the set @xmath61 must be a crossing tile . this means observation [ o : corner ] applies to some row or column of @xmath12 , and @xmath12 can be reduced to case 3 . hence @xmath12 does not represent @xmath26 . + by the above four cases , we see that the @xmath26 can not be placed on a @xmath20 mosaic . hence , by the figure for @xmath26 in the appendix , we see that the mosaic number of @xmath26 is six ; that is @xmath27 . we next consider the seven - crossing knots . as seen above , the knot @xmath47 can be placed on a @xmath20 mosaic board . we formalize this result in the following proposition as well as establish the mosaic number for the other seven - crossing knots . + [ t:7 - 4 ] the mosaic number of @xmath47 is five , that is @xmath62 . moreover , @xmath47 is the only seven - crossing prime knot with mosaic number five ; the remaining seven - crossing prime knots have mosaic number six . we have already seen via observation [ o : corner ] that at most seven crossing tiles can be placed on a @xmath63 board without reduction via a reidemeister type i move . moreover , since all seven crossing knots are alternating , there is only one way to place seven alternating crossing tiles up to mirror , reflection , and rotation as depicted in figure [ f : alt_7 ] . when this arrangement is suitably connected , the only knot resulting is @xmath47 . therefore @xmath62 and all other seven - crossing knots have mosaic number six as depicted in the appendix . next we consider the eight - crossing knots . let @xmath3 be a knot of eight crossings . by the proof of theorem [ t:6 - 3 ] , we know that @xmath3 can not fit on a @xmath20 mosaic board . this means @xmath64 . furthermore there exists a knot mosaic of @xmath3 on a @xmath25 board ( see the appendix ) . this implies @xmath65 . + given the above arguments , we see that the mosaic number of the eight crossing knots is greater than five . by the appendix and use of knotscape , we see that the mosaic number of all eight crossing knots is six . we summarize our findings in the table [ t : obvious ] . for each knot @xmath3 with at most 8 crossings , the appendix includes a mosaic of size @xmath5 representing @xmath3 . + 10 colin c. adams , bevin m. brennan , deborah l. greilsheimer , and alexander k. woo , _ stick numbers and composition of knots and links _ , j. knot theory ramifications , * 6 * ( 1997 ) , no . 2 , 149161 . colin c. adams , _ the knot book _ , american mathematical society , providence , ri , 2004 . j. w. alexander _ a lemma on systems of knotted curves _ , usa * 9 * ( 1923 ) 9395 . knotscape , http://www.math.utk.edu/~morwen/knotscape.html . kyungpyo hong , ho lee , hwa jeong lee , and seungsang oh , _ small knot mosaics and partition matrices _ , j. phys . a , * 47 * ( 2014 ) , no . 43 , 13 pp . t. kuriya and o. shebab , _ the lomonaco - kauffman conjecture _ , j. knot theory ramifications , * 23 * ( 2014 ) , no . 1 hwa jeong lee , kyungpyo hong , ho lee , and seungsang oh , _ mosaic number of knots _ , j. knot theory ramifications , * 23 * ( 2014 ) no . 13 , 8 pp . samuel j. lomonaco and l. kauffman , _ quantum knots and mosaics _ , quantum inf . , * 7 * ( 2008 ) , no . 2 - 3 , 85115 . lewis d. ludwig , erica l. evans , and joseph s. paat , _ an infinite family of knots whose mosaic number is realized in non - reduced projections _ , j. knot theory ramifications , * 22 * ( 2013 ) , no . 7 , 11 pp . kurt reidemeister , _ elementare begrdung der knotentheorie _ , hamburg , * 5 * ( 1926 ) , 2432 .	in 2008 , lomonaco and kauffman introduced a knot mosaic system to define a quantum knot system . a quantum knot is used to describe a physical quantum system such as the topology or status of vortexing that occurs on a small scale can not see . kuriya and shehab proved that knot mosaic type is a complete invariant of tame knots . in this article , we consider the mosaic number of a knot which is a natural and fundamental knot invariant defined in the knot mosaic system . we determine the mosaic number for all eight - crossing or fewer prime knots . this work is written at an introductory level to encourage other undergraduates to understand and explore this topic . no prior of knot theory is assumed or required . future planning ( nrf-2015r1c1a2a01054607 ) .
the profiles of some pulsars include the component usually called interpulse ( ip ) , which is separated from the main pulse ( mp ) by approximately a half of the pulsar period . the ips may be as weak as a few per cent of the mp and typically have much steeper spectra , being most pronounced at low frequencies @xcite . in several pulsars , the ip is connected to the mp with a barely distinguishable emission bridge , and the low - level emission may extend over practically the whole pulse period ( e.g. * ? ? ? * ) . in general , the ip is located not exactly midway between the successive mps , but may be shifted by up to a few tens degrees toward earlier or later pulse phases . in case of still larger shifts the ip is rather referred to as the postcursor or the precursor to the mp ( e.g. * ? ? ? * ; * ? ? ? * ; * ? ? ? note that the pulse profile may contain a few such components at once . for example , the crab pulsar has a total of six components outside of the mp , which exhibit pronounced frequency evolution @xcite . the emission components outside of the mp are characteristic of the short - period pulsars , @xmath0 s @xcite , and especially of the millisecond pulsars @xcite . the ips are met in @xmath1 of the normal pulsars and @xmath2 of the millisecond ones . as a rule , the ip separation from the mp does not change with frequency @xcite . this contrasts with the spectral behaviour of the component separations inside the mp , which are known to increase with wavelength , signifying the overall broadening of the emission cone . besides that , the ips usually have distinctive polarization characteristics , showing higher precentage of linear polarizaton and a shallow position angle swing , the position angle itself being not very much different from that of the mp . although the classical model of rotating vector explains successfully the position angle swing across the mp in a number of pulsars , its application to the profiles with ips often faces difficulties ( e.g. * ? ? ? the single - pulse observations further reveal the peculiarities of the ip emission . the ip intensity appears modulated at various timescales . similarly to the mps , the ips may exhibit a number of fluctuation phenomena , such as microstructure , subpulse drift , pulse - to - pulse intensity modulation , mode changing , and giant pulses . in psr b0950 + 08 , the ip shows microstructure comparable with that in the mp @xcite . recently it has been found that the subpulse patten in the ip of psr b1702 - 19 has the same periodicities as that in the mp , and moreover , the two patterns are intrinsically in phase @xcite . this striking result directly testifies to the physical relation between the mp and ip emissions . some indirect manifestations of such a relation have presumably been known previously . in particular , in psr b0950 + 08 , a strong mp is followed by a strong ip @xcite , whereas in psr b1055 - 52 a strong ip is followed by a strong mp @xcite . in both cases the strong components appear separated by more than a half of the pulse period . it is still obscure whether this peculiar intensity correlation is an artifact of the subpulse drift or not . apart from the weak intensity fluctuations of the ip emission similar to those in the mp , occasional strong pulses , with intensities as large as about the mp intensity , can be met in these pulsars at the position of the ip @xcite . these transient events are so rare that neither the average ip profile nor the above mentioned mp - ip correlation are affected . in psr b1822 - 09 , the ip shows quite different fluctuation behaviour , participating in mode changing @xcite . the ip appears pronounced only in the so - called quiet mode , i.e. in the sequence of weak enough pulses . in the bright mode , a strong precursor arises some @xmath3 ahead of the mp , whereas the ip becomes markedly weaker . thus , the ip intensity is anticorrelated with both the mp and precursor intensities . provided that the giant pulse activity is characteristic of the mp ( e.g. in the crab pulsar and psr b1937 + 21 ) , giant pulses can also be met in the ip , though they are not necessarily simultaneous in the two components at a given frequency . in the crab pulsar , giant mps and ips are found to exhibit quite distinct temporal and frequency structures @xcite . giant mps consist of one or several broadband microbursts made up of shorter and narrowband nanoshots ( @xmath4 , @xmath5 ) . giant ips contain microsecond - long trains of proportionally spaced emission bands ( @xmath6 ) , which are grouped into regular band sets . in summary , the ip emission is characterized by a number of peculiarities , and at the same time shows diversiform manifestations of its physical connection to the mp . all this calls for theoretical explanation . in the preceding literature , the ips are interpreted in terms of several geometrical models . one of the models suggests that the ip is emitted from the magnetic pole opposite to that responsible for the mp and the magnetic axis is nearly orthogonal to the rotational axis , in which case the emission from both poles is alternately seen by an observer . this two - pole model can satisfactorily explain the main features of the ip components , except for their physical relation to the mp emission and the continuous emission bridge , which may connect them to the mp . these difficulties are removed in the single - pole models . in the first version of such a model , the mp and ip are identified with the two components of a hollow emission cone under the condition of unusually large angular extent of the cone and/or approximate alignment of the magnetic and rotational axes of a pulsar @xcite . although this model naturally explains the emission bridge between the mp and the ip as well as the physical connection of the components , it is not clear why the components so much differ in intensity and their separation does not change with frequency like that between the conal components of the ordinary narrow profiles . later on @xcite attempted to improve the single - pole model in order to avoid these difficulties . it has been assumed that the two concentric hollow cones , the inner and outer ones , are centered at the magnetic axis , which is almost aligned with the rotational axis . then the portions of the inner and outer emission cones grazed by the sight line form the mp and the ip in the resultant pulse profile . recently @xcite have suggested a bidirectional model of pulsar emission and applied it to the peculiar profile of psr b1822 - 09 . in that model , the mp and the precursor originate independently at different altitudes above the same magnetic pole and the precursor emission intermittently reverses its direction to form the ip . this so - called inward emission directed toward the neutron star can be observable provided that the magnetic and rotational axes are nearly orthogonal . the physics underlying the reversal of the emission direction is not specified , but for any conceivable switching mechanism it is difficult to explain the connection of the emission direction to the mp intensity . in the present paper , we suggest the first physical model of ips . in contrast to the above mentioned geometrical models , it is aimed at explaining the spectral and polarization peculiarities of the ip emission as well as the nature of the mp - ip connection . recently we have proposed the physical mechanism of the precursors @xcite . it is based on induced scattering of the mp emission into the background . it has been found that in case of a superstrong magnetic field the scattered radiation is directed almost along the field . then , because of rotational aberration in the scattering region , the scattered component appears in the pulse profile up to @xmath7 ahead of the mp . in the present paper , we extend the theory of magnetized induced scattering to the case of a moderately strong magnetic field . it will be shown that in this approximation the mp emission is scattered in the direction antiparallel to the ambient magnetic field and may form the profile component roughly midway between the mps . the plan of the paper is as follows . in section 2 , we examine the problem on the radio beam scattering into the background in application to pulsars . the properties of the scattered component are compared with the observed features of the ip emission in section 3 . a summary of our model of the ip formation is given in section 4 . basic formalism of induced scattering in the approximation of moderately strong magnetic field is given in appendix . pulsar magnetospheres contain the ultrarelativistic electron - positron plasma , which streams along the open magnetic lines . the radio emission is believed to originate deep inside the open field line tube , and therefore it should propagate through the plasma flow . as the brightness temperatures of pulsar radio emission are extremely high , the induced scattering off the plasma particles may be substantial . the induced scattering of radio emission by the non - magnetized pulsar wind has been considered by @xcite . however , inside the magnetosphere of a pulsar , the magnetic field may be strong enough to affect the scattering process considerably . this happens on condition that the radio frequency in the particle rest frame is much less than the electron gyrofrequency , @xmath8 ( here @xmath9 is the particle lorentz - factor , @xmath10 the velocity in units of @xmath11 , @xmath12 is the angle the incident photon makes with the particle velocity , and the quantities not denoted by primes correspond to the laboratory frame ) . the magnetized scattering has been studied in a number of papers ( e.g. * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) . in the vicinity of the emission region of a pulsar , the regime of magnetized scattering is certainly valid . as the magnetic field strength decreases with distance from the neutron star , @xmath13 , at high enough altitudes @xmath14 , i.e. the cyclotron resonance takes place . the resonance region typically lies in the outer magnetosphere , and hence the regime of magnetized scattering , @xmath15 , holds over a substantial part of the open field line tube well below the light cylinder . the pulsar radio emission is believed to be generated at the frequencies of order of the local lorentz - shifted proper plasma frequency , @xmath16 , where @xmath17 and @xmath18 is the plasma number density ( but see * ? ? ? * for the criticism of this point ) . hence , in the emission region and its close neighborhood , the scattering is the collective process . the induced scattering of different types of the plasma waves is an important ingredient of various scenarios of the pulsar emission mechanism ( e.g. * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? besides that , the resultant radio waves may participate in the induced three - wave interactions @xcite . as the plasma number density decreases with the distance from the neutron star , @xmath19 , well above the emission region @xmath20 , in which case the plasma effects are negligible and the induced scattering is a single - particle process ( e.g. * ? ? ? * ) . in the present paper , we dwell on the magnetized induced scattering in the single - particle approximation . then the incident radiation presents the transverse waves linearly polarized either in the plane of the ambient magnetic field or perpendicularly to this plane . actually , there are two regimes of magnetized scattering , the longitudinal and transverse ones ( e.g. * ? ? ? * ; * ? ? ? as the strength of the external magnetic field tends to infinity , the excited motion of a particle in the field of the incident wave is confined to the magnetic field line . this is a so - called longitudinal scattering . because of a purely longitudinal motion of the particle in this regime , only the photon states with the polarization in the plane of the ambient magnetic field are involved in the scattering . in case of somewhat weaker magnetic fields , the excited motion of a particle presents a drift in the crossed fields , the electric field of the incident wave and the ambient magnetic field . in case of large enough transverse component of the perturbed particle velocity the character of the scattering changes substantially . in particular , this so - called transverse scattering involves the photons of both polarizations , with the electric vectors in the plane of the magnetic field and perpendicular to this plane . in case of spontaneous scattering of the radiation directed at the angle @xmath21 to the magnetic field , the longitudinal regime holds if @xmath22 , whereas the transverse one on condition that @xmath23 ( e.g. * ? ? ? for the induced scattering of radio beam into background , which will be studied in the present paper , the condition of switching between the regimes is expected to be somewhat modified , and this question will be examined in detail in section 2.5 . at the conditions relevant to pulsar magnetosphere , both regimes are believed to be appropriate . previous studies of magnetized induced scattering in application to pulsars have concentrated on the longitudinal regime @xcite . this process appears efficient , and it may cause a number of observational consequences . in particular , induced longitudinal scattering of the radio beam into the background may account for the low - frequency turnovers in pulsar spectra @xcite , whereas the scattered component may be identified with the precursor to the mp @xcite ; the scattering inside the beam results in the photon focusing , which may underlie the formation of microstructure in pulsar profiles @xcite ; induced scattering between the two beams of substantially different frequencies and orientations leads to significant intensity redistribution in frequency and can explain giant pulses along with their nanostructure @xcite . the induced scattering in the transverse regime has been briefly discussed in @xcite . general formalism of this process is developed in the appendix of the present paper . the kinetic equations obtained are used to solve the problem of the radio beam scattering into the background . the main motivation of our study is that in contrast to the longitudinal scattering regime , the radiation is believed to be scattered backwards , in the direction antiparallel to the particle velocity . therefore we suggest the induced transverse scattering as a mechanism underlying the formation of interpulses . let the radio beam pass through the open field line tube of a pulsar and be scattered off the particles of the plasma flow . the radiation of pulsars is known to be highly directional : at any point of the pulsar emission cone it is concentrated into a narrow beam with the opening angle @xmath24 , whereas the width of the emission cone itself ( which determines the observed pulse width ) is typically much larger . far from the emission region , the radiation propagates quasi - transversely with respect to the ambient magnetic field , @xmath25 . therefore one can neglect the beam width and represent the incident radiation by a single wavevector @xmath26 at any point of the scattering region . ( note that at different locations within the pulsar emission cone the wavevector orientations somewhat differ , so that the observed pulse has a non - zero width . ) the rate of induced scattering is known to depend on the particle recoil , i.e. on the difference of the initial and final directions of the photons , and therefore the scattering within the beam is of no interest here . on the other hand , by definition the induced scattering can not transfer the beam photons into the states where the photon occupation numbers are initially zero . however , the photons can still be subject to induced scattering out of the beam , since some background photons are expected to be always present in space . in particular , they may result from the spontaneous scattering of the beam , which , in contrast to the induced one , provides the photons of any orientation . although in pulsar case the spontaneous scattering is very inefficient and the scattered photons are too scanty to be detectable ( see , e.g. , eq.[13 ] below ) , they can still stimulate a substantial induced scattering from the radio beam . the beam photons should predominantly undergo induced scattering into the background state @xmath27 , which corresponds to the maximum scattering probability . if the induced scattering is efficient enough , the background radiation in this state may grow significantly and become almost as strong as the initial radio beam . it should be noted that in the particle rest frame the photon frequency is approximately unchanged in the scattering act . therefore in the laboratory frame @xmath28 . thus , we examine the problem on induced scattering between the two photon states , @xmath29 and @xmath30 , one of which corresponds to the beam and another one to the background state characterized by the maximum scattering probability ; the frequencies are related as @xmath31 , and the occupation number of the background photons is initially much less than that of the beam photons , @xmath32 . the kinetic equations describing the rate of the magnetized induced scattering are derived in appendix ( see eqs.[a7 ] ) . to proceed further we introduce several simplifications . first of all , as we concentrate on the transverse scattering , only the last terms in the kinetic equations ( a7 ) should be retained . then these equations can be presented as @xmath33 where @xmath34 @xmath35 stand for the polarization states of the incident and scattered photons , and @xmath36 one can see that the kinetic equations differ from each other only by the factor @xmath37 , which is generally of order unity . ( note the symmetry of @xmath37 with respect to the initial and final photon states ) . further , a detailed form of the particle distribution function does not play a crucial role . therefore we consider a monoenergetic distribution with some characteristic lorentz - factor of the particles . it is convenient to replace the photon occupation numbers with the intensities , @xmath38 and @xmath39 , and making use of their delta - functional angular distributions to integrate the kinetic equation over the solid angle . then we come to the following system of equations for the spectral intensities of the beams @xmath40 : @xmath41 where @xmath42 and @xmath43 is the particle number density . obviously , the radio beam intensity @xmath44 decreases because of the photon scattering to the state with @xmath45 and the maximum scattering probability corresponds to @xmath46 and @xmath47 . note that in this situation the azimuthal angle @xmath48 is of no interest . the system ( 4 ) has the following solution ( see , e.g. , * ? ? ? * ) : @xmath49 where @xmath50 is the first integral . thus , in our approximate consideration the induced scattering results in the net intensity transfer of the radio beam intensity into the background . of course , an exact treatment of the problem taking into account the complete rather than approximate cross - sections would show that the total intensity @xmath51 somewhat decreases , the energy being deposited to the scattering particles , and the number of photons @xmath52 is conserved instead of the intensity @xmath53 . one can see that as a significant fraction of photons comes to the background state , @xmath54 , the corresponding intensity is @xmath55 . hence , of the original energy of the radio beam , @xmath56 , about @xmath57 is transferred to the background state and @xmath58 is gained by the particles . although @xmath59 , the background intensity grows drastically ( cf . eq.[13 ] below ) , and the intensity transfer between the two states greatly dominates the evolution of the total intensity @xmath51 . in our treatment , the latter is ignored and the intensity of the efficiently growing component is intended to be roughly comparable with the original radio beam intensity or at least be above the detection level . according to equation ( 6 ) , the efficiency of intensity transfer is characterized by the quantity @xmath60 . provided that @xmath61 the background intensity grows exponentially , @xmath62 , but still remains much less than the original radio beam intensity , and , correspondingly , @xmath63 . a significant part of the beam intensity is transferred to the background on a more stringent condition : @xmath64 at still larger @xmath65 , the background intensity increases very weakly , slowly approaching the initial radio beam intensity , @xmath66 , whereas the beam intensity decreases considerably , @xmath67 . it is expected that in pulsar case @xmath68 , i.e. the scattered component is strong enough to be observable , while the beam is not suppressed drastically . to have a notion about the efficiency of transverse scattering in pulsar magnetosphere let us first estimate the quantity @xmath69 , where @xmath70 is given by equation ( 5 ) at @xmath71 and @xmath72 . the original intensity is related to the radio luminosity of a pulsar , @xmath73 , as @xmath74 where @xmath75 is the pulse width in the angular measure , @xmath76 the spectral index , and it is taken that @xmath77 hz . the number density of the scattering particles can be expressed in terms of the goldreich - julian density , @xmath78 where @xmath79 is the plasma multiplicity factor , and @xmath80 the pulsar period . since @xmath81 , @xmath19 , and @xmath82 , the main contribution to the scattering depth @xmath83 comes from the region near the cyclotron resonance , where @xmath84 . the radius of cyclotron resonance can be estimated as @xmath85^{1/3},\ ] ] where @xmath86 is the magnetic field strength at the surface of the neutron star and all the quantities are normalized to their typical values . making use of equations ( 9)-(11 ) , we obtain the scattering efficiency : @xmath87 to conclude if this is sufficient to satisfy the inequality ( 8) let us estimate the initial level of the background radiation resulting from the spontaneous transverse scattering of the radio beam : @xmath88 . substituting the scattering cross - section in the form @xmath89 , which is roughly applicable at @xmath90 , and taking @xmath91 and @xmath92 , we find that @xmath93 given that @xmath94 , @xmath95 is necessary for the background intensity to increase appreciably as a result of the induced transverse scattering . obviously , at the conditions relevant to pulsar magnetosphere , @xmath83 can indeed be as large as a few tens . it should be noted that the original radio luminosity entering @xmath96 may noticeably exceed the average radio luminosity deduced from observations because of pulse - to - pulse fluctuations of the radio emission and also because of intensity suppression in the course of radio beam scattering into the background . according to equation ( 12 ) , the scattering appears more efficient for larger original luminosities , shorter periods , and lower frequencies . this is in line with the observational statistics . it is interesting to compare the efficiencies of induced scattering into the background in the transverse and longitudinal regimes . as can be seen from equation ( a7a ) , for a fixed @xmath97 the term corresponding to the transverse scattering dominates on condition that @xmath98 . but it is necessary to take into account that the transverse scattering is the strongest at @xmath46 , whereas the longitudinal one peaks at @xmath99 . comparing the quantity @xmath70 given by equation ( 5 ) at @xmath100 with that given by equation ( 9 ) in @xcite , we find that the maximum scattering efficiencies in the two regimes are related as @xmath101 thus , at frequencies @xmath90 the transverse scattering is much more efficient . deeper in the magnetosphere , however , the ambient magnetic field is much stronger , the incident intensity and the particle number density entering @xmath102 are larger , so that the longitudinal scattering may dominate . in general , the processes of induced scattering in the two regimes are expected to compete in efficiency and can both be significant . however , if @xmath103 even in the emission region , the longitudinal scattering is inefficient at all . note that the emission altitude is not known accurately , especially for the millisecond pulsars . besides that , in the vicinity of the emission region the angle @xmath12 between the wavevector and the ambient magnetic field is also uncertain , since it may be determined by refraction rather than by the magnetosphere rotation . hence , it is not clear whether the regime of longitudinal scattering is the case in all pulsars over the whole radio frequency range . as for the transverse scattering , it occurs anyway . it is worthy to mention here that , according to equation ( 11 ) , for the parameters of the pulsars which exhibit ips , the characteristic altitude of transverse scattering is of order of the light cylinder radius . one can speculate that this is the necessary condition for the pulsar to show the ip . now let us consider the location of the scattered radiation in the pulse profile . for the sake of simplification it is assumed that the emission altitude is negligible as compared to the altitude of the scattering region , the ray geometry is dominated by the effect of magnetosphere rotation at the point of scattering , and the magnetic axis of a pulsar is perpendicular to the rotational axis . the scattered component resulting from the transverse scattering is antiparallel to the ambient magnetic field , and therefore is expected to be identified with the ip component in the pulse profile . the ray @xmath26 emitted approximately along the magnetic axis at @xmath104 at the point o ( see fig . 1 ) comes to the point of scattering s at @xmath105 , while the magnetic axis turns by the angle @xmath106 . the polar angle of the point of scattering with respect to the instantaneous magnetic axis is @xmath107 , and in the dipolar geometry the angle between the ambient magnetic field vector @xmath108 and the magnetic axis is @xmath109 . hence , the angle between @xmath26 and @xmath108 is @xmath110 . in the corotating frame , the wavevector of the scattered radiation , @xmath111 , is nearly antiparallel to @xmath108 , whereas in the laboratory frame it is shifted because of aberration by the angle @xmath107 in the direction of rotation . as can be seen from fig . 1 , the scattered ray @xmath112 travels from the point of scattering to the point i for @xmath113 and later on reaches the observer . the main pulse is characterized by the parallel ray @xmath114 , which is emitted along the magnetic axis and points toward the observer . the ray @xmath115 originates later than @xmath26 by @xmath116 , where @xmath117 is the angle between the two instantaneous positions of the magnetic axis , and comes to the point m for the time @xmath118 . here it should be noted that the points i and m are equidistant from the point o , but not from the observer . the point i is located @xmath119 $ ] farther , the distinction becoming significant at @xmath120 . taking into account that the difference in pulse longitude is related to the difference in arrival times as @xmath121 , we find finally that @xmath122 or , equivalently , @xmath123 thus , the ip lags the mp by more than @xmath124 . figure 2 shows the separation of the two components as a function of the height of the scattering region . for a number of pulsars , the ip positions in the pulse profile appear compatible with our model . at the same time , it is difficult to explain the cases when the ip lags the mp by less than @xmath124 . note that our geometrical examination is based solely on the effect of rotational aberration in the scattering region . a more realistic consideration including the magnetic field distortions close to the light cylinder and the propagation effects on the scattered radiation would possibly release the requirement of a nearly orthogonal rotator and allow the scattered component to lag the mp by less than @xmath125 to account for the postcursors in the pulsar profiles . the induced compton scattering of pulsar radio emission off the secondary plasma particles inside the open field line tube may play a significant role . the magnetic field of a pulsar affects the scattering process substantially . at distances of order of the emission altitude , the magnetic field is typically strong enough for the scattering to occur in the longitudinal regime , @xmath126 . then the photons of the pulsar radio beam are predominantly scattered into the state with @xmath127 and @xmath128 , i.e. the scattered radiation is almost aligned with the ambient magnetic field . as the magnetic field strength decreases with distance from the neutron star , at higher altitudes the longitudinal scattering regime changes for the transverse one , which holds on condition that @xmath129 . in this regime , the orientations of the scattered photons are mostly antiparallel to the magnetic field , @xmath46 , and @xmath47 . the radiation scattered in the two regimes may from separate components in the pulse profile , which can be identified with the precursor and the ip . apart from explaining the geometrical location of these components in the pulse profile , our model suggests the mp - ip and mp - precursor connections as well as can account for the peculiar properties of the emission components outside of the mp . in the geometrical aspect , our model is akin to the bidirectional model of @xcite . in that model , the precursor and ip originate at a certain location in the outer magnetosphere and the pulsar is approximately an orthogonal rotator . because of the ultrarelativistic outstream of the plasma particles along the field lines , any conceivable emission mechanism would generate the outward radiation directed along the magnetic field in the corotating frame . the inward radiation is assumed to be antiparallel to the outward one . thus , the orientations of the emission components and their locations in the pulse profile are roughly the same as in our model . note , however , that in our case the precursor and the ip result from the scattering in the two distinct regimes and therefore originate at somewhat different altitudes . besides that , the scattering sites are restricted to the region of the radio beam passage in the rotating magnetosphere . correspondingly , the positions of the precursor and the ip in the pulse profile are tightly connected to that of the mp . as the scattered component arises in the outer magnetosphere , its position in the pulse profile is determined by the rotational aberration . the scattering region lies at distances of order of the cyclotron resonance radius , @xmath130 , where @xmath131 . taking into account that @xmath132 and @xmath133 one can see that @xmath130 is a weak function of radio frequency , @xmath134 , and hence the ip separation from the mp is almost the same over a wide frequency range . this is compatible with observations : as a rule , the ip position in the profile appears independent of frequency ( e.g. * ? ? ? * ) . note that in the simplified case of orthogonal rotator and purely dipolar magnetic field considered in the present paper the rotational aberraion causes the ip to lag the mp by more than a half of the pulsar period . a more general treatment of the pulsar geometry including the peculiarities close to the light cylinder , is expected to account for smaller ip separations as well . in the approximation considered , the induced scattering results in a net intensity transfer from the radio beam to the background , the total intensity of the two beams being constant . hence , the maximum intensity of the scattered component , @xmath135 , is restricted to the original intensity of the radio beam , @xmath136 . since the intensity is transferred to the lower frequency , @xmath137 , and the pulsar radiation has a decreasing spectrum , @xmath138 may appear considerably less than the mp intensity at the same frequency . in a number of cases , the ips are indeed much weaker than the mp . it should be noted , however , that at a fixed frequency the mp and ip may compete in intensity if the mp is noticeably suppressed by the scattering into still lower frequencies . we have examined the scattering of radiation of a fixed frequency , but , generally speaking , it is a broadband process . since the scattering efficiency @xmath139 strongly depends on frequency , the scattering should affect the spectra of the mp and ip emissions . according to equation ( 12 ) , at lower frequencies the scattering is more efficient and , although the intensity transfer approaches the stage of saturation , @xmath68 , the growth of the ip component is believed to be noticeably stronger . this trend agrees with the observations , which testify that the ip phenomenon is most pronounced at the decameter wavelengths , at the edge of the observed radio frequency range @xcite . furthermore , the scattering can markedly suppress the mp emission at low frequencies , so that the mp spectrum may somewhat flatten . the ips really have steeper spectra than the mps , with the intensity ratio of the ip and mp markedly decreasing with frequency ( e.g. * ? ? ? the estimate of the scattering efficiency ( 12 ) implies that the intensity transfer is more significant at larger radio luminosities , shorter periods , and weaker magnetic field strengths . the ips are indeed met in the pulsars with periods @xmath0 s @xcite . moreover , the population of normal short - period pulsars , @xmath140 s , is generally characterized by larger radio luminosities than that of the long - period ones . as for the millisecond pulsars , their luminosities are somewhat less @xcite , but very short periods , @xmath141 ms , and weak magnetic fields , @xmath142 g , favor even larger scattering efficiencies . note that the pulsars with ips are indeed more abundant in the population of the millisecond pulsars . our model of ip formation as a result of induced scattering of the mp implies peculiar polarization properties of the scattered component . in contrast to the longitudinal scattering , when the intensity may be efficiently transferred only between the photon states with the ordinary polarization and the scattered component is characterized by the complete linear polarization , the transverse scattering involves both orthogonal polarizations and the situation is more complicated . for different channels of the scattering , the efficiency of intensity transfer differs by the factor @xmath37 , which is generally of order unity . in case of intense scattering ( see eq.[8 ] ) , the difference in @xmath37 and @xmath143 for various channels may play a significant role , and the intensity transfer in one of the channels may substantially dominate that in the others , so that the scattered component may be strongly polarized . note that the observed ips are typically characterized by higher percentage of linear polarization than the mps ( e.g. * ? ? ? * ; * ? ? ? * ) , with the giant ips showing almost complete linear polarization @xcite . our model also suggests a specific behavior of the position angle of linear polarization in the ip emission . the position angle of the scattered radiation is determined by the orientation of the @xmath144-plane in the scattering region and therefore should somewhat differ from that of the mp . besides that , the mp and ip may be dominated by different polarization modes , in which case the position angle of the ip is additionally shifted by @xmath145 . as the scattering region lies in the outer magnetosphere , in the area covered by the radio beam the magnetic field is almost uniform and hence the position angle should remain practically unchanged across the ip . all this is in line with the observational data ( e.g. * ? ? ? * ; * ? ? ? * ; * ? ? ? as is discussed above , the origin of ips as a result of the mp scattering far from the emission region can account for a number of distinctions in the properties of the mp and ip emissions . at the same time , our model implies a physical connection between these emissions , which is believed to manifest itself as a correlation of the temporal fluctuations of the mp and ip and also as a consistency of the angular and frequency structures in the two components . the idea of the mp - ip connection is strongly supported by a number of observational results . recent observations of psr b1702 - 19 @xcite have shown that the subpulse pattern in the ip is characterized by exactly the same periodicities as that in the mp . moreover , the intensity fluctuations appear correlated with a delay of about a half of the pulsar period . this is just what can be expected if the mp is partially scattered into the ip . as the subpulse pattern is independent of frequency , the subpulse modulation in the scattered component should repeat that in the incident radiation . the mp and the ip seen by an observer originate at different phases of pulsar rotation , and therefore they should arise at different phases of subpulse drift and the intensities should be correlated with a certain temporal delay . it is worth noting that this delay should not exactly correspond to the longitudinal separation between the mp and ip in the profile , since the components travel somewhat different distances to the observer . the microstructure characteristic of the mp emission is also expected to be present in the scattered component . the observations of psr b0950 + 08 have indeed revealed the microstructure in the ip at the timescale @xmath146s , whereas in the mp @xmath147s @xcite . in our model , the relationship between the microstructure timescales in the two components can be estimated as follows . the intensity is transferred between the photon states related as @xmath148 . differentiating this at fixed frequencies yields @xmath149 , where it is taken that @xmath150 for @xmath151 . as @xmath152 and @xmath153 , one can find that @xmath154 . taking into account that the microstructure timescale evolves with frequency , @xmath155 , we obtain finally : @xmath156 in case of psr b0950 + 08 , @xmath157 s and @xmath147s , so that the angular scale of microstructure @xmath158 . then with @xmath159 and @xmath160 we have @xmath161 , which is consistent with the observed value of about 0.7 . thus , our model can account for the difference in the microstructure timescales of the mp and the ip . recent observations of the mp and ip of the crab pulsar @xcite have revealed quite distinct temporal and frequency structures of the giant pulses in these components . the giant mps generally present one or several broadband microbursts , which consist of narrowband ( @xmath5 ) nanoshots of a duration @xmath162 s. the giant ips consist of the proportionally spaced narrow emission bands of microsecond lengths organized into several band sets , which appear at somewhat different times and exhibit a marked drift toward higher frequencies . below we examine the modification of the temporal and frequency structure of the giant mps as a result of induced scattering into the background and argue that the consequent structure of the scattered component is compatible with that really seen in the giant ips . since the intensity is transferred between the frequencies approximately related as @xmath163 , the frequency structure of the ip at a fixed pulse longitude can be derived from differentiating this equation : @xmath164 . as the nanoshot width is extremely small , @xmath165 ( here the period of the crab pulsar @xmath166 s ) , the main contribution to @xmath167 comes from the first term , and we have @xmath168 thus , the proportionally spaced emission bands of the ip(with @xmath169 ) can naturally be attributed to the modification of the nanoshot bandwidths as a result of induced scattering . the temporal durations of the bands in the ip can be estimated by means of equation ( 16 ) . note that because of the extremely short durations of the original nanoshots , the scattered nanoshots are noticeably enlarged : @xmath170 . for @xmath160 , @xmath171 and @xmath172 s we have @xmath173s , which agrees with the observational data . keeping in mind that @xmath174 , one can see that at a given frequency @xmath175 different parts of the nanoshot and the neighboring nanoshots are scattered to somewhat different @xmath176 , leading to a slight drift of the scattered bands toward higher frequencies , just as is really observed . thus , the peculiar temporal and frequency structure of the giant ips can indeed be explained in terms of induced scattering of the giant mps . the pulsar b1822 - 09 is known to exhibit a peculiar mode changing behavior : in the bright mode , its profile contains a strong precursor and a weak ip , whereas in the weak mode the ip is strong and the precursor is almost absent @xcite . within the framework of our model , this can be interpreted as a competition between the processes of induced scattering of the mp in the longitudinal and transverse regimes . the variations of the scattering efficiencies can naturally be attributed to the fluctuations of the number density of the scattering particles . furthermore , the fluctuations of the plasma number density lead to changes in the emission altitude of the mp radiation , which may affect the applicability of the longitudinal scattering regime . larger multiplicities of the plasma , @xmath79 , imply larger emission altitudes for a given frequency , in which case the gyrofrequency appears low enough to preclude the longitudinal scattering even in the vicinity of the emission region . at the same time , larger @xmath79 favor stronger transverse scattering . as a result , the precursor component is not formed , whereas the intensity transfer to the ip is so efficient that the mp is markedly suppressed . as the intensity is transferred from higher to lower frequencies , @xmath177 , and the radio beam originally has a decreasing spectrum , the ip remains weak as compared to the mp at the same frequency . thus , the weak mode is characterized by larger plasma multiplicities and , correspondingly , by an efficient intensity transfer from higher to lower frequencies , in which case the total intensity of the pulse profile markedly decreases . in the bright mode , the plasma number density is smaller , the emission altitude lower , and the longitudinal scattering holds in addition to the transverse one . the longitudinal scattering gives rise to the precursor in the pulse profile , whereas the transverse scattering , being less efficient because of lower @xmath79 , forms a weaker ip . as is evident from equation ( 14 ) , the transverse scattering may be much more efficient in taking the intensity from the mp at a given frequency . at the same time , if one compare the intensity growth of the precursor and the ip at a given frequency , it is necessary to keep in mind that these components are fed by the mp intensity at substantially different frequencies , @xmath178 and @xmath179 . with the decreasing spectrum of pulsar radiation , this implies much larger @xmath180 than that given by equation ( 14 ) , so that both scattering efficiencies can be concurrently substantial . since the longitudinal scattering transfers the intensity to the precursor from much lower frequencies , this component may be strong enough as compared to the mp at the same frequency . note that the precursor in the profile of psr b1822 - 09 is really comparable in intensity to the mp and is much stronger than the ip , providing a strong support to our scenario of intensity transfer between widely spaced frequencies in the course of induced scattering of the mp into the background . we have considered the induced compton scattering of radio emission by the particles of the hot magnetized electron - positron plasma , which streams along the open field lines of the pulsar magnetosphere . in the presence of a strong magnetic field , @xmath181 , two scattering regimes are possible , the longitudinal and transverse ones , both being relevant to the pulsar case . the earlier studies of the magnetized induced scattering in pulsars were typically restricted to the longitudinal regime . in the present paper , we have generalized the kinetic equation to include the transverse scattering regime as well and have used it to solve the problem on the transverse induced scattering of the radio beam into the background . our problem can be reduced to examining the scattering between the two photon states , one of which represents the radio beam , whereas another one is the background state corresponding to the maximum scattering probability for the beam photons . since in the particle rest frame the photon frequency is almost unchanged in the scattering act , in the laboratory frame the photon states interacting via induced scattering obey the condition @xmath182 . in contrast to the longitudinal regime , in which case the radio beam photons are predominantly scattered into the direction nearly along the ambient magnetic field ( @xmath127 and @xmath183 ) , the photons scattered in the transverse regime are mostly antiparallel to the field ( @xmath184 and @xmath185 ) . in both regimes , the induced scattering results in the intensity redistribution between the two photon states , with the total intensity remaining approximately unchanged . this specific character of intensity evolution is determined by the role of the magnetic field in the scattering process and differs substantially from the non - magnetic case , when the scattering results in the photon drift toward lower frequencies . the magnetized induced scattering of the radio beam into the background , which takes place in the open field line tube of a pulsar , is suggested to underlie the formation of additional components in the pulse profile outside of the mp . given that the pulsar is a nearly orthogonal rotator , the longitudinal scattering gives rise to the precursor component located a few tens degrees ahead of the mp , whereas the transverse scattering forms the ip , which lags the mp by more than @xmath125 . the numerical estimates show that at the conditions relevant to pulsar magnetospheres the intensity transfer from the radio beam to the background can indeed be efficient . stronger scattering is favored by shorter pulse periods , larger radio luminosities and lower frequencies and is expected to be especially efficient in the millisecond pulsars . all this is in line with the trends known for the observed ip emission . note that the ip component is fed by the mp radiation of much higher frequencies , @xmath186 . with the decreasing spectrum of the original mp , this implies that the ip can not be as large as the mp at the same frequency unless the latter is substantially suppressed by the scattering to still lower frequencies . our model can account for the peculiar properties of the ip emission . as the region of efficient transverse scattering lies far from the emission region , at distances of order of the cyclotron resonance radius , and the characteristic scattering altitude is an extremely weak function of frequency , the mp - ip separation should remain almost the same over the observed radio frequency range . note that the scattering at somewhat lower altitudes may also be noticeable , giving rise to the emission bridge between the mp and the ip . the position angle of linear polarization of the scattered radiation is determined by the orientation of the @xmath187-plane in the scattering region and generally differs from the position angle of the incident radiation . furthermore , the magnetic field is believed to be almost uniform throughout the scattering region , so that the position angle swing in the ip should be shallow . although both orthogonal polarization modes can grow as a result of the transverse scattering , the efficiencies of intensity transfer may differ markedly , making expect high percentage of linear polarization in the resultant ip emission . the ip formation as a result of the mp scattering implies a physical connection between these components , which is believed to manifest itself in their intensity fluctuations at different timescales . our model is strongly supported by the recent observations of psr b1702 - 19 @xcite , which have revealed that the subpulse patterns in the mp and the ip are intrinsically in phase . furthermore , the peculiar temporal and frequency structure of the giant pulses in the ip of the crab pulsar @xcite may well be interpreted as a modification of the giant mp structure in the scattering process . the peculiar moding behavior of psr b1822 - 09 is suggested to result from the fluctuations of the physical conditions in the scattering region , which affect the relative efficiency of the longitudinal and transverse scatterings and lead to the observed interplay between the resultant precursor and ip components . the pulse - to - pulse fluctuations of the plasma number density and the original mp intensity , which affect the scattering efficiency , are believed to result in the peculiar intensity statistics of the ip . moreover , in some pulsars this component may be seen only occasionally . note that the transient ips are difficult to detect and they are yet to be found observationally , though some evidence for occasional strong events at the ip longitudes is already known @xcite . despite the relative weakness of the ips in most of the known cases , the present progress in the observational facilities seems to promise an opportunity of the comprehensive study of the ip emission at a single - pulse level . in the present paper , we for the first time suggest a physical model of the ip formation . it is believed to stimulate further observational investigations aimed at revealing the peculiar properties of the ip emission and the manifestations of its connection to the mp . i am grateful to the anonymous referee for useful suggestions and criticisms . following @xcite , let us consider the induced scattering in the laboratory frame between the two photon states , @xmath188 and @xmath27 , involving the electrons with the initial momenta @xmath189 and @xmath190 . the electrons are confined to move along the magnetic field line , and in the scattering act the momentum parallel to the field is conserved , so that @xmath191 where @xmath12 and @xmath97 are the wavevector tilts to the magnetic field for the photons in the states @xmath188 and @xmath192 . the rate of change of the photon occupation number as a result of induced scattering is @xmath193\frac{{\rm d}p}{{\rm d}t}n_1{\rm d}p.\ ] ] here @xmath194 and @xmath195 are the photon occupation numbers in the states @xmath188 and @xmath196 , respectively , @xmath197 is the electron distribution function , @xmath198 is the probability of generating spontaneously scattered photons per electron per unit time , @xmath199 where @xmath200 , @xmath201 , @xmath202 is the scattering cross - section per unit solid angle , and the argument of the delta - function means that in the particle rest frame the initial and final frequencies are equal , @xmath203 . we assume that the photon occupation numbers change only because of the photon propagation through the stationary flow , so that @xmath204 . taking into account that @xmath205 , we integrate equation ( a2 ) by parts , substitute equations ( a1 ) and ( a3 ) and integrate this over @xmath206 with the help of delta - function to obtain @xmath207{\rm d}\omega_1.\ ] ] here @xmath208 stands for the distribution function in lorentz - factor . the above kinetic equation describes the photon transfer as a result of induced scattering in a hot magnetized plasma . we consider the scattering of the ordinary and extraordinary transverse electromagnetic waves . their polarization states , with the electric vectors in the plane of the wavevector and the ambient magnetic field and orthogonal to this plane , are designated as a- and b - polarizations , respectively . the classical scattering cross - section in an arbitrary magnetic field has been derived by @xcite . expanding their result in a power series of @xmath209 and retaining the first two terms yield here the superscripts of @xmath214 denote the initial and final polarization states of a photon , the primes mark the quantities in the electron rest frame , @xmath215 is the classical electron radius , @xmath216 , @xmath217 and @xmath218 are the spherical angles of the initial and final photon wavevectors @xmath219 and @xmath220 in the coordinate system with the polar axis along the magnetic field , and @xmath221 . using relativistic transformations , @xmath222 one can express the cross - sections in terms of the quantities of the laboratory frame , substitute this into equation ( a4 ) and perform differentiation with respect to @xmath9 . it should be noted that the dominant term of the cross - sections , @xmath223 as well as the remaining factor in the expression under differentiation in equation ( a4 ) depend on @xmath9 only implicitly , via @xmath224 and @xmath225 , which are weak functions of @xmath9 : @xmath226 . therefore it is necessary to retain the second - order terms , @xmath227 , which introduce the explicit dependence on @xmath9 . although they are small , their derivatives may contribute significantly . keeping in mind these considerations , one can obtain the kinetic equations in the following form : @xmath228 \right\}{\rm d}\omega_1\nonumber\\ + \frac{\hbar nr_e^2}{mc}\int { \rm d } \gamma f\int\frac{\omega^2}{\omega_g^2}\sin^2\delta\phi\left \{\frac{(\eta-\eta_1)^2}{\beta^2\gamma^3\eta_1 ^ 2}\frac{1}{k } \frac{\partial n_1k_1 ^ 3}{\partial k_1}\frac{(1-\eta\gamma^2)^2}{\beta^2\gamma^4\eta^2 } \frac{(1-\eta_1\gamma^2)^2}{\beta^2\gamma^4\eta_1 ^ 2}\right.\nonumber \\ \left . + \frac{2(\eta-\eta_1)kn_1\eta^3}{\beta^6\gamma^{11}\eta_1 ^ 3}\left [ \frac{(\eta\gamma^2 - 1)[\eta\gamma^2(2-\eta)-1]}{\eta^3 } \frac{(1-\eta_1\gamma^2)^2}{\eta_1 ^ 2 } \right.\right.\nonumber \\ \left.\left . + \frac{(\eta_1\gamma^2 - 1)[\eta_1\gamma^2(2-\eta_1)-1]}{\eta_1 ^ 3 } \frac{(1-\eta\gamma^2)^2}{\eta^2}\right ] \right\}{\rm d}\omega_1\nonumber \\ -\frac{\hbar nr_e^2}{mc}\int { \rm d } \gamma f\int\frac{\omega^2}{\omega_g^2}\frac{\cos\delta\phi\sin 2\theta\sin 2\theta_1}{\mu\mu_1\beta^2\gamma^3}\left \{\frac{(\eta-\eta_1)^2}{2\gamma^2\eta_1 ^ 2}\frac{1}{k } \frac{\partial n_1k_1 ^ 3}{\partial k_1}\right.\nonumber \\ \left . -\frac{(\eta-\eta_1)kn_1\eta^3}{\eta_1 ^ 3}\left [ 1-\frac{\mu + \mu_1 } { 2\gamma^2\mu\mu_1}\right ] \right\}{\rm d}\omega_1\nonumber \\ + \frac{\hbar nr_e^2}{mc}\int{\rm d}\gamma f\int\frac{\omega^4}{\omega_g^4 } \frac{(\eta -\eta_1)n_1k\eta^4(\eta\gamma^2 - 1)}{\eta_1 ^ 3\beta^2\gamma } \left [ 2(1+\sin^2\delta\phi)\frac{(1-\eta\gamma^2)^2 } { \beta^2\gamma^4\eta^2}\frac{(1-\eta_1\gamma^2)^2}{\beta^2\gamma^4\eta_1 ^ 2}\right.\nonumber \\ \left . -\cos\delta\phi\sin 2\theta\sin 2\theta_1 \right ] { \rm d}\omega_1 .\end{aligned}\ ] ] @xmath229\right \}{\rm d}\omega_1 \nonumber \\ + \frac{2\hbar nr_e^2}{mc}\int { \rm d } \gamma f\int\frac{\omega^4}{\omega_g^4}(1+\cos^2\delta\phi)\frac{(1-\eta_1\gamma^2)^2 } { \beta^2\gamma^4\eta_1 ^ 2}\frac{(\eta -\eta_1)n_1k\eta^4(\eta\gamma^2 - 1)}{\eta_1 ^ 3\beta^2\gamma}{\rm d}\omega_1,\\ \frac{{\rm d}n}{{\rm d}r}(a\to b)=\frac{\hbar nr_e^2}{mc}\int { \rm d } \gamma f\int\frac{\omega^2}{\omega_g^2}\cos^2\delta\phi \left \{\frac{(\eta-\eta_1)^2}{\beta^2\gamma^3\eta_1 ^ 2}\frac{1}{k } \frac{\partial n_1k_1 ^ 3}{\partial k_1}\frac{(1-\eta\gamma^2)^2}{\beta^2\gamma^4\eta^2 } \right.\nonumber \\\left . + \frac{(\eta-\eta_1)kn_1\eta^3}{\eta_1 ^ 3}\frac{2(\eta\gamma^2 - 1 ) } { \beta^4\gamma^7\eta^3}[\eta\gamma^2(2-\eta)-1]\right \}{\rm d}\omega_1 \nonumber \\ + \frac{2\hbar nr_e^2}{mc}\int { \rm d } \gamma f\int\frac{\omega^4}{\omega_g^4}(1+\cos^2\delta\phi)\frac{(1-\eta\gamma^2)^2 } { \beta^2\gamma^4\eta^2}\frac{(\eta -\eta_1)n_1k\eta^4(\eta\gamma^2 - 1)}{\eta_1 ^ 3\beta^2\gamma}{\rm d}\omega_1.\end{aligned}\ ] ] the first term of the kinetic equation ( a7d ) is determined by the photon spectrum and is qualitatively similar to the right - hand side of the kinetic equation for the non - magnetic case ( cf . , e.g. , eq.[2.13 ] in @xcite ) , signifying the monotonic shift of the photon distribution toward lower frequencies in the course of the scattering . the second term means the redistribution of photons between the states which satisfy the condition @xmath231 . because of the factor @xmath232 , the photon occupation numbers decrease as a result of the photon transfer to the states with @xmath233 and increase on account of the photons coming from the states with @xmath234 . one can find that the ratio of the second term in equation ( a7d ) to the first one is @xmath235 , where @xmath236 . thus , in the case of interest , in a moderately strong magnetic field , the second term dominates . although the kinetic equations ( a7b ) and ( a7c ) are somewhat more complicated , their second terms also dominate on the same condition . the equation ( a7a ) is worthy to be analyzed in more detail . its first term , corresponding to the first term of the cross - section ( a5a ) , does not contain the gyrofrequency and describes the longitudinal scattering , which remains efficient at @xmath237 . the second item of this term differs from the first one by the factor @xmath238 and hence dominates at @xmath239 . in the regime of transverse scattering , @xmath240 , the last term of equation ( a7a ) dominates , being at least a factor of @xmath241 larger than the first one and @xmath242 larger than the second and the third ones . the sign of the integrand in the last term is determined by that of @xmath243 , that is the photons are transferred to the states with @xmath233 . this is similar to the corresponding terms in the kinetic equations ( a7b ) , ( a7c ) , and ( a7d ) and contrasts with the longitudinal regime , in which case the photons are transferred closer to the magnetic field direction , @xmath244 ( cf . the second item in the braces of the first term in equation ( a7a ) ; for more details see @xcite ) . eilek , j. a. , & hankins , t. h. 2007 , in proceedings of the 363 . we - heraeus seminar : neutron stars and pulsars ( posters and contributed talks ) physikzentrum bad honnef , germany , may.14 - 19 , 2006 , eds . w.becker , h.h.huang , mpe report 291 , p.112	the magnetized induced compton scattering off the particles of the ultrarelativistic electron - positron plasma of pulsar is considered . the main attention is paid to the transverse regime of the scattering , which holds in a moderately strong magnetic field . we specifically examine the problem on induced transverse scattering of the radio beam into the background , which takes place in the open field line tube of a pulsar . in this case , the radiation is predominantly scattered backwards and the scattered component may grow considerably . based on this effect , we for the first time suggest a physical explanation of the interpulse emission observed in the profiles of some pulsars . our model can naturally account for the peculiar spectral and polarization properties of the interpulses . furthermore , it implies a specific connection of the interpulse to the main pulse , which may reveal itself in the consistent intensity fluctuations of the components at different timescales . diverse observational manifestations of this connection , including the moding behavior of psr b1822 - 09 , the peculiar temporal and frequency structure of the giant interpulses in the crab pulsar , and the intrinsic phase correspondence of the subpulse patterns in the main pulse and the interpulse of psr b1702 - 19 , are discussed in detail . it is also argued that the pulse - to - pulse fluctuations of the scattering efficiency may lead to strong variability of the interpulse , which is yet to be studied observationally . in particular , some pulsars may exhibit transient interpulses , i.e. the scattered component may be detectable only occasionally .
damped ly@xmath0 ( dla ) absorbers are the class of qso intervening systems with highest hydrogen column density , @xmath3 atoms @xmath4 . several properties of these absorbers indicate that they are high - redshift progenitors of present - day galaxies ( wolfe et al . 1986 , wolfe et al . spectroscopic observations show , in addition to high hi column densities , metal lines of low ionization and velocity dispersions typical of gas rich galaxies . imaging studies of fields around qsos with foreground dla absorption at low redshift ( @xmath5 ) reveal the presence of gas - rich galaxies with a variety of morphological types ( le brun et al . an indepedent and remarkable piece of evidence in favour of a galactic origin comes from the behaviour of @xmath6 , the comoving mass density of neutral gas in dla absorbers in units of the critical density of the universe . the fact that @xmath7 equals the mass density of luminous matter ( stars ) at the present epoch , and that @xmath6 decreases with time , is interpreted as evidence of gas consumption due to star formation since @xmath8 to the present time ( wolfe et al . 1995 ) . at high redshift @xmath6 is at least 10% of the barionic content of the universe and therefore understanding the nature of dla galaxies will give important contraints to cosmological models involving structure formation in the early universe ( klypin et al . a commonly accepted origin of dla absorbers is in spiral galaxies , ( wolfe et al . 1995 , prochaska & wolfe 1997b ) , but an origin in dwarf galaxies has also been suggested ( york et al . 1986 ; tyson 1988 ; pettini , boksenberg & hunstead 1990 ; molaro et al . 1996 ; matteucci , molaro & vladilo 1997 ) . studies of elemental abundances in dla systems give important clues for understanding which kind(s ) of galaxies are responsible for the absorption . in addition , the study of dla abundances over a large redshift interval will eventually allow us to trace the chemical evolution of the absorbers over a large interval of look - back time , up to the early stages of galactic nucleosynthesis , providing fundamental tests to models of galactic chemical evolution . abundance measurements in dla systems do not depend critically on ionization corrections because these absorbers are optically thick to ionizing radiation ( h@xmath9 ev ) . as a consequence , species that are dominant in hi regions have negligible ionization corrections , as confirmed by detailed photoionization calculations ( lu et al . 1995 ; viegas 1995 ) . nevertheless , the abundances observed in dla absorbers may not represent the real chemical composition of the system if part of the elements is removed from the gas to the solid phase ( dust grains ) , as it happens in the interstellar medium of our galaxy where elements with very different dust depletions are known to exist ( jenkins 1987 ; savage & sembach 1996 ) . in dla absorbers the relative abundances of elements expected to have different dust depletions , such as zn and cr , suggest that dust is indeed present , even though with a lower dust - to - gas ratio than in the galaxy ( pettini et al . 1994 , 1997 ) . more direct evidence for dust in dla systems has been found from reddening studies of qsos with and without foreground damped absorption ( pei , fall & bechtold , 1991 ) . a relatively large number of abundance determinations in dla absorbers has been collected since it has become possible to observe qsos at sufficiently high spectral resolution ( see list of references quoted in table 2 ) . attempts to interpret the relative abundances ( i.e. element - to - element abundance ratios ) observed in dla absorbers have lead , however , to contradictory conclusions . from one side claims have been made that the pattern of relative abundances resembles that of galactic halo stars , having in common an enhancement of the alpha elements and a deficiency of manganese relative to iron ( lu et al . 1996 , prochaska & wolfe 1997 ) . from the other side , evidence has been presented for nearly solar [ @xmath0/fe ] ratios in spite of the low metallicity level of the absorbers ( molaro et al . 1996 , molaro et al . these contradictory conclusions are probably due to the fact that the presence of dust in dla absorbers is not properly taken into account . several attempts have been done to reproduce the observed abundances by combining different intrinsic abundance patterns with dust depletion patterns of galactic interstellar type ( lauroesch et al . 1996 ; kulkarni , fall & truran 1997 ) . according to these studies , neither a halo - like nor a solar - like metallicity pattern match the observations perfectly . in the present work the effect of the dust on the determinations of the elemental abundances is studied in detail by deriving equations that take into account the level of metallicity , the dust - to - gas ratio and the intrinsic abundance pattern of dla absorbers . differently from the analysis performed by kulkarni , fall & truran ( 1997 ) , in the present study each individual dla system is treated separately and the extinction induced by the dust is taken explicitely into account . the general equations that link observed abundances to real abundances of the absorbers are presented in section 2 . simplified forms of these equations are then used in section 3 to predict possible dust depletion patterns in dla systems . in section 4 is presented the method for estimating dust - to - metals ratios , dust - to - gas ratios and dust - corrected abundances in individual dla absorbers ; the method is applied to available abundance measurements of dla systems . the relations between dust and metals in dla absorbers are examined in section 5 . the implications for our understanding of the nature of dla galaxies are discussed in section 6 and the results are summarized in section 7 . the observed abundance ratio by number ( x / h)@xmath10 of an element x in a dla absorber is determined from measurements of the column densities in the gas phase of the system , @xmath11 and @xmath12 . the observed abundance is usually expressed in logarithmic form relative to solar composition @xmath13_{{\rm obs}}= \log ( { { { \rm x}}\over { { \rm h } } } ) _ { { \rm obs}}- \log ( { { { \rm x}}\over { { \rm h } } } ) _ { \sun } ~.\ ] ] however , this is not the real abundance since the column density of atoms in dust , @xmath14 , is not considered in ( 1 ) . the total column density ( @xmath15 ) would instead give the real ratio ( x / h)@xmath16 and hence the real abundance @xmath13_{{\rm dla}}= \log ( { { { \rm x}}\over { { \rm h } } } ) _ { { \rm dla}}- \log ( { { { \rm x}}\over { { \rm h } } } ) _ { \sun } ~.\ ] ] by introducing the _ fraction in dust _ of the element x @xmath17 it is easy to derive a relation between the observed and the real abundance @xmath13_{{\rm obs}}= \left [ { { { \rm x}}\over { { \rm h } } } \right]_{{\rm dla}}+ \log \left ( 1 - f_{{{\rm x } } } \right ) ~~.\ ] ] to introduce the effect of the dust in the above expression we start by defining the dust - to - gas ratio @xmath18 where @xmath19 is the extinction optical depth in the spectral region of interest in the rest frame of the absorber . the above definition is the same of pei , fall & bechtold ( 1991 ) provided the hydrogen column density@xmath20 is expressed in units of 10@xmath21 @xmath4 . if we call @xmath22 the column density of dust grains per unit area along the line of sight , the optical depth is given by @xmath23 where @xmath24 is the _ average extinction cross - section of the grains_. for spherical grains of radius @xmath25 the cross - section for extinction is @xmath26 where the efficiency factor for extinction @xmath27 can be estimated in the framework of mie scattering theory ( see e.g. wickramasinghe 1967 ) . if we call @xmath28 the number of grains with radii in the range ( @xmath25,@xmath29 ) per unit volume , then it is possible to define @xmath30 through the relation @xmath31 we will use , however , the most general form ( 6 ) to avoid any assumption on the geometry of the grains . we then define the _ average number of atoms of the element x per dust grain _ @xmath32 which is related to the chemical composition of the dust since the number ratio of two elements x and y in the dust grains will be ( x / y)@xmath33 = @xmath34 . from ( 5 ) , ( 6 ) and ( 7 ) it is easy to see that @xmath35 , where @xmath36 . by defining the metallicity level of the system as @xmath37 , where y is a suitable reference element , we obtain : @xmath38 from ( 4 ) and ( 8) we obtain a relation between observed and real abundance : @xmath13_{{\rm obs}}= \left [ { { { \rm x}}\over { { \rm h } } } \right]_{{\rm dla}}+ \log \left ( 1- { k \over z } \left ( { { { \rm x}}\over { { \rm y } } } \right)^{-1}_{{\rm dla } } { \nu_{{\rm x}}\over c_{\rm e } } \right)\ ] ] where the contributions of dust - to - gas ratio , metallicity level , elemental abundance pattern , dust chemical composition and dust extinction are clearly separated . a crucial role in the relation is played by the _ dust - to - metals ratio _ @xmath39 . relation ( 9 ) is not suited for comparing observed and real abundances because it includes several unknown quantities . in order to obtain a more useful relation it is convenient to normalize the quantities in the dla absorbers in units of the corresponding galactic quantities . we indicate such normalized values with a tilded symbol , e.g. @xmath40 . by applying ( 8) to dla systems and to our galaxy , we derive @xmath41_{{\rm dla } } } \cdot { { \tilde{\nu}}_{{\rm x}}\over { \tilde{c}}_{\rm e } } \cdot f_{{{\rm x}},{{\rm ism } } } ~,\ ] ] where ( i ) @xmath42_{{\rm dla}}}$ ] , ( ii ) @xmath43_{{\rm dla } } } = ( { { \rm x}}/{{\rm y}})_{{\rm dla}}/ ( { { \rm x}}/{{\rm y}})_{\sun}$ ] since we adopt solar abundances for the galaxy , and ( iii ) the fraction in dust @xmath44 refers to the galactic interstellar medium . this expression for the fractions in dust can be written in a simpler form if we require the condition @xmath45 for any element , which is equivalent to assume that _ the dust has same chemical composition and same average number of atoms per grain as galactic dust . _ this equivalence follows from the fact that the relative abundance of two elements in the dust is @xmath46 . a constant relative abundance in the dust is consistent with current understanding of grains chemical composition ( see e.g. discussion in section 12.2 and 12.3 by savage & sembach 1996 ) . in practice , once the chemical composition and the average number of atoms are specified , also the grain sizes are specified . therefore , also the extinction coefficient , which depends on the dielectric properties and geometry of the grains is determined , the only left degree of freedom being the exact shape of the grains . therefore , once the condition @xmath45 is satisfied , also condition @xmath47 is virtually satisfied , and hence @xmath48 . relation ( 10 ) becomes @xmath49_{{\rm dla } } } \cdot f_{{{\rm x}},{{\rm ism}}}\ ] ] and from ( 4 ) we obtain the equation @xmath13_{{\rm obs}}= \left [ { { { \rm x}}\over { { \rm h } } } \right]_{{\rm dla}}+ \log \left ( 1 - { { \tilde{k}}\over { \tilde{z } } } \cdot 10^ { - \left [ { { { \rm x}}\over { { \rm y } } } \right]_{{\rm dla } } } \cdot f_{{{\rm x}},{{\rm ism } } } \right)\ ] ] that can be used to estimate the gas - phase abundances given the real metallicity pattern and a suitable combination of dust - to - gas ratio @xmath50 , and metallicity @xmath51 . the fractions in dust @xmath44 can be considered as input parameters which are known with reasonable accuracy from galactic interstellar studies , as we show in the next section . by analogy to the definition of logarithmic dust depletion in the galactic interstellar medium @xmath52 we define the dust depletion in dla systems as @xmath53 where the number ratios @xmath54 are measured in the gas phase in each case . since dust depletion and fraction in dust ( 3 ) are related by the equation @xmath55 it follows from ( 11 ) that for galactic - type dust @xmath56_{{\rm dla } } } \cdot f_{{{\rm x}},{{\rm ism } } } \right ) .\ ] ] we have used this relation to estimate dust depletion patterns in dla systems by using the dust - to - metals ratio @xmath57 and the abundance pattern @xmath58_{{\rm dla}}$ ] as free parameters . the results are shown in table 1 , where elements most commonly measured in dla systems are considered . iron is used as the zero point for the metallicity level . we have computed the fractions in dust @xmath44 from ( 15 ) by adopting dust depletions typical of galactic warm disk ( savage & sembach 1996 ) . more details are given in the notes to table 1 . the results in the columns labeled s100 , s75 , and s50 are relative to a solar abundance pattern ( @xmath59_{{\rm dla}}= 0 $ ] ) for dust - to - metals ratios @xmath57 = 1.00 , 0.75 and 0.50 , respectively . by definition , the case s100 is nothing else than the galactic depletion pattern for low density lines of sight ( galactic warm disk ) . by comparing this pattern with the cases s75 and s50 one can see that the absolute values of the depletions decrease when the dust - to - metals ratio is reduced . however , _ the scaling law of the differential depletion is not linear _ , contrary to what assumed in previous works ( see e.g. lauroesch et al . rather , _ the differential depletion between pairs of elements tends to vanish as the dust - to - metals ratio decreases_. this implies that elements moderately depleted in the galaxy can still have a non negligible depletion when less dust is present . as an example , si depletion decreases by only 0.3 dex , compared to a decrease of 1 dex of fe , when @xmath57 is reduced from 1.0 to 0.5 . more surprising results are found when an abundance pattern different from the solar one is considered . in the columns labeled h100 , h75 , and h50 we give the depletions expected for an intrinsic abundance pattern typical of galactic halo stars , again for dust - to - metals ratios @xmath57 = 1.00 , 0.75 and 0.50 , respectively . the adopted values of halo star abundances are taken from francois ( 1988 ) , sneden , gratton & crocker ( 1991 ) , and ryan et al . ( 1996 ) , at a metallicity level @xmath60 characteristic of dla absorbers ; we refer to the notes in table 1 for more details . the very small deviations from the solar pattern found for [ cr / fe ] and [ ni / fe ] are treated here as real to show how even small deviations can affect the dust depletion pattern . from the results shown in table 1 one can see that _ inversions in the differential depletion pattern may appear if the dla abundance pattern is not solar . _ for instance , underabundant elements such as mn and cr are more depleted than fe in halo - like models , contrary to what found in galactic interstellar gas . overabundant elements tend , on the other hand , to be less depleted . as an example , ti depletion is quite modest in halo - like models while it is comparable to iron depletion in galactic gas . it is also interesting to note that _ no solutions are allowed for underabundant elements with high fraction in dust ( @xmath61 ) unless the dust - to - metals ratio is sufficiently low . _ this apparently peculiar result is due to the fact that one can not use _ whatever _ combination of input parameters because by definition ( 3 ) must be @xmath62 and therefore @xmath63_{{\rm dla } } } / f_{{{\rm x}},{{\rm ism } } } ) $ ] , otherwise there will be not enough atoms available to form the required amount of dust . as an example , there are no solutions for mn in the models h100 and h75 , while there is solution for h50 , when the dust - to - metals ratio becomes half the galactic value . instead of using eq . ( 12 ) to estimate the effect of a given dla abundance pattern on the observed abundances , we will follow the opposite approach of deriving the real abundance pattern starting from the observed ratios . by applying eq . ( 12 ) to the elements x and y and combining the resulting equations it is possible to find the real abundance ratio @xmath64_{{\rm dla } } } = 10^ { \left [ { { { \rm x}}\over { { \rm y } } } \right]_{{\rm obs } } } + { { \tilde{k}}\over { \tilde{z } } } \cdot \left ( f_{{{\rm x}},{{\rm ism } } } - f_{{{\rm y}},{{\rm ism } } } \cdot 10^ { \left [ { { { \rm x}}\over { { \rm y } } } \right]_{{\rm obs } } } \right)\ ] ] from the observed abundance ratio . this equation still depends on the dust - to - metals ratio which is unknown . however , if two elements x and y trace each other over the chemical history of the system , i.e. if @xmath65_{{\rm dla}}= 0 $ ] at any time , then it is easy to derive from ( 17 ) the expression @xmath66_{{\rm obs } } } - 1 \over f_{{{\rm y}},{{\rm ism } } } \cdot 10^ { \left [ { { { \rm x}}\over { { \rm y } } } \right]_{{\rm obs } } } - f_{{{\rm x}},{{\rm ism } } } } \ ] ] that allows to compute the dust - to - metals ratio from observed abundances only , by using the galactic interstellar fractions in dust as input parameters . once the dust - to - metals content is determined from a given pair of elements , eq . ( 17 ) can be applied to other elements for deriving their intrinsic abundance . this method can be applied to individual dla systems for which suitable abundances measurements are available . in the following sub - sections we describe the results of the application of this procedure to currently available measurements . in order to determine the dust - to - metals ratio from eq . ( 18 ) we need at least two elements for which @xmath65_{{\rm dla}}= 0 $ ] in dla systems . studies of metal poor stars in the galaxy indicate that iron - peak elements elements generally trace each other down to metallicities comparable to those found in dla absorbers ( see third column of table 1 and references quoted therein ) . a relatively large amount of observations of iron - peak elements are available for dla systems , as can be seen in the updated summary of measurements presented in table 2 . the use of iron - peak elements is therefore the most natural choice to determine the dust - to - metals ratio . however , a few caveats have to be considered . from one side , observed dla [ mn / fe ] ratios show an underabundance which is difficult to understand in terms of dust depletion ( lu et al . in addition , some deviations from solar ratios are found in metal - poor stars for [ cr / fe ] , [ mn / fe ] , [ co / fe ] ( ryan et al . 1996 ) , and for [ cu / fe ] ( sneden , gratton & crocker 1991 ) . for cr , however , these deviations are significant only at [ fe / h ] @xmath67 and are negligible at the metallicity level typical of dla systems . taking also into account the number of available measurements , this leaves zn , fe and cr as the best candidates for using eq . ( thanks to the long - standing observational effort of pettini et al . ( 1994 , 1997 ) , a relatively large number of zn and cr measurements are available . these overlap in many cases with fe ( and sometimes ni ) determinations based on keck observations ( lu et al . 1996 and refs . therein ) . the form of the denominator of the right side of ( 18 ) indicates that the use of any pair of elements with similar fractions in dust and gas - phase abundances must be avoided , because a small observational error would lead to a large error in the dust - to - metals ratio . this precludes the use of cr and fe , as can be seen in tables 1 and 2 . since zn has a much smaller fraction in dust than cr and fe we have determined the dust - to - metals ratio from ( 18 ) by using zn as the element x and cr or fe as the element y. table 2 includes the 17 dla systems for which a zn measurement exists , rather than an upper limit , together with a cr or fe determination . the resulting dust - to - metals ratios are shown in table 3 . in 8 cases the determinations are obtained independently from [ zn / cr ] and [ zn / fe ] ratios . we do not find a systematic difference between the 8 results in common , the average percent difference being @xmath68 . this agreement gives us confidence on the method adopted and on the assumption that these elements indeed trace each other in dla systems , at least down to the level of metallicity of the present sample . for these 8 cases we adopt the average of the two independent determinations as the final dust - to - metals ratio . for the other cases we adopt the individual result , being confident that fe vs. cr systematic effects should be negligible . an attempt to derive the dust - to - metals ratio using [ zn / ni ] leads to values systematically higher by about 20% than the values derived from [ zn / cr ] and [ zn / fe ] . this suggests that a weak deviation from solar iron - peak relative composition may be present for ni , even though the suspect has been advanced that ni abundances could be affected by uncertainties in the niii oscillator strengths ( lu et al . 1996 ) . for these reasons , dust - to - metals ratios determined from [ zn / ni ] ratios are not considered here . the typical random error of the dust - to - metals ratio derived in this way can be estimated by neglecting the contribution of @xmath69 in the denominator of the right side of ( 18 ) . the error of the dust - to - metals ratio equals that of the [ zn / y ] measurement , which is typically of 0.1 dex , although in some cases it can be as low as 0.05 dex . in addition to this random error , systematic errors are introduced through the galactic fractions in dust used in ( 18 ) . to quantify this effect we computed the dust - to - metals in two cases : ( i ) zinc completely undepleted , i.e. @xmath70 , and ( ii ) zinc mildly depleted , with @xmath71 . the resulting dust - to - metals ratios , also shown in table 3 , are systematically lower by @xmath72 20% in the first case relative to the second case . uncertainties in fe and cr fractions in dust would yield a much smaller systematic error since @xmath73 for current range of fe and cr dust depletions observed in galactic interstellar clouds . having derived the @xmath74 ratio for individual dla systems we can now use relation ( 17 ) to estimate relative abundances corrected from dust effects . in table 4 we list additional abundance measurements available for the dla systems of the sample shown in table 3 . dust - corrected relative abundances are given in table 4 below each gas - phase determination and are shown in a synoptic view in fig . 1 . internal errors of the corrected relative abundances are quite low , as estimated from the dispersion of the 5 [ si / fe ] measurements , @xmath75 dex , and of the 7 [ ni / fe ] measurements , @xmath76 dex . by applying error propagation to ( 17 ) one can see that the total random error of the dust - corrected ratios is typically @xmath72 0.15 dex , but can be as low as @xmath72 0.05 dex for the most precise measurements . in addition , systematic errors related to the uncertainty of the @xmath44 and @xmath77 parameters should be taken into account . the use of dust - to - metals ratios derived by assuming zinc completely undepleted would systematically reduce the dust corrections ; however , the net effect on the corrected ratios is in most cases less than 0.05 dex . a first result clear from fig . 1 is that the spread of the dust - corrected relative abundances for any given [ x / fe ] ratio is generally within the range of the expected errors , suggesting that _ the sample of dla systems investigated is chemically homogeneous over the range of metallicities from nearly solar down to 1.5 dex . _ the most important result visible if fig . 1 , however , comes from the analysis of the element - to - element abundance pattern . if the intrinsic abundances of dla systems were similar to metal - poor galactic halo stars , the ratios shown in fig . 1 should deviate from the solar pattern , with the only exception of [ ni / fe ] . quite surprisingly _ the dust - corrected ratios are remarkably close to solar . _ errors due to uncertainties of the fractions in dust @xmath44 and @xmath77 should produce different effects for any given pair of elements x and y , and would tend to destroy the solar - like pattern . of course , as we discuss in detail in the following paragraphs , the relative abundances are not _ exactly _ solar . it is also important to note here that departures from solar ratios in dla systems are found when nitrogen abundances are considered ( lu 1997 , priv . comm . ; vladilo et al . however , owing to the complex galactic chemical evolution of this element , the study of nitrogen abundances is deferred to a subsequent paper ( centurin et al . we discuss now in more detail specific ( groups of ) elements . nickel is expected to trace iron at least down to a metallicity @xmath78 , even though there is some indication for a very mild underabundance ( @xmath79 dex ) at @xmath80 from the data by ryan et al . the 7 dust - corrected measurements available yield @xmath81 > = -0.06 \pm 0.05 $ ] , consistent both with ni tracing fe or with a mild underabundance . even if this result is not particularly surprising , it shows already the importance of the correction from dust effects . in fact the gas phase [ ni / fe ] measurements yield values around @xmath82 dex , which have lead lu et al . ( 1996 ) to suspect the presence of a constant offset between stellar and dla measurements . our results can be instead explained in terms of the expected trends of iron - peak elements . the ratios of alpha elements to iron - peak elements is the best indicator of halo - like metallicity pattern . in the early stages of chemical evolution the products of type ii snae , rich in alpha elements , are expected to dominate the observed abundances and to produce the overabundance of the [ @xmath0/fe ] ratio which is observed in galactic metal - poor stars . at later stages , the contribution of type ia snae reduces the [ @xmath0/fe ] ratio to the typical values observed in galactic stars of solar metallicity ( see e.g. matteucci , molaro & vladilo 1997 ) . at a metallicity @xmath80 galactic halo stars show [ si / fe ] = + 0.27 dex ( ryan et al . 1996 ) and [ s / fe ] = + 0.4 dex ( francois 1988 ; including a correction of @xmath82 dex proposed by lambert 1989 ) . ti behaves similarly to an @xmath0 element with [ ti / fe ] = + 0.22 at @xmath80 ( ryan et al . if the chemical history of dla systems were similar to that of the galaxy , we would expect to find similar overabundances of the [ @xmath0/fe ] ratios . this is not the case for any of the elements considered here . the 5 dust - corrected measurements of si yield @xmath83 > = + 0.02 \pm 0.09 $ ] ( internal error of the average ) . even assuming a total random error as high as 0.15 dex , the result is still inconsistent at two sigma level from the [ si / fe ] @xmath72 + 0.3 dex expected for a halo - like abundance pattern . also the 3 dust - corrected measurements of [ s / fe ] do not show evidence of overabundance . as noted by molaro , centurin & vladilo ( 1997 ) , in all these 3 cases the observed [ s / zn ] ratio is solar . since dust depletion is modest for both s and zn this result confirms the validity of the correction applied to the observed [ s / fe ] ratios . of the two dust - corrected measurements of [ ti / fe ] , one is clearly not overabundant and the other is consistent , within an error bar of @xmath84 dex both with a solar - like abundance and with the mild overabundance observed in galactic halo stars . finally , we note that also the [ mg / fe ] upper limit derived by molaro , centurin & vladilo ( 1997 ) for the @xmath85 absorbers toward phl 957 is inconsistent with a halo - like pattern . gas - phase measurements of the [ mn / fe ] ratio in dla systems systematically yield negative values ( lu et al . the [ mn / fe ] ratio is believed to be a good discriminant between galactic halo and galactic interstellar abundance patterns . in the first case mn is underabundant ( [ mn / fe ] = @xmath86 dex ; ryan et al . 1996 ) while in latter case mn is overabundant ( [ mn / fe ] @xmath72 @xmath87 ; savage & sembach 1996 ) as a consequence of differential dust depletion . however , as we have shown in section 3 , even a mild underabundance of the [ mn / fe ] ratio can lead to an inversion of the dust depletion pattern , manganese becoming more depleted than iron . this explains why the dust correction can increase the observed [ mn / fe ] ratio , as it happens in the two cases shown in table 4 . the dust corrected ratios indicate in fact a modest underabundance of manganese . this underabundance is smaller than expected for a halo - like chemical composition and the [ mn / fe ] ratio is marginally consistent with a solar composition . we note that this conclusion does not depend on the adopted cosmic abundance for manganese . for consistency with the results of ryan et al . ( 1996 ) we have renormalized the dla measurements by adopting solar photospheric abundance for mn , instead of the solar meteoric value used by lu et al . ( should we adopt the meteoritic value , we would find a [ mn / fe ] lower by 0.14 dex in the dla systems , but also the [ mn / fe ] ratio for galactic halo stars should be decreased , down to [ mn / fe ] = @xmath88 dex , and the dust corrected ratios would still be in between the solar - like and the halo - like case . the basic assumption of the present work is that dust affects the observed abundances in dla absorbers . we can test this assumption by comparing the dust - to - gas ratios derived from the observed abundances with dust - to - gas ratios obtained from reddening determinations . for any two elements x and y with similar nucleosynthetic history it is possible to estimate the dust - to - gas ratio by means of the expression @xmath89_{{\rm obs } } } - 10^ { \left [ { { { \rm x}}\over { { \rm h } } } \right]_{{\rm obs } } } \over f_{{{\rm x}},{{\rm ism } } } - f_{{{\rm y}},{{\rm ism } } } } \ ] ] which can be derived by applying ( 12 ) to x and y and by combining the resulting relations with the condition @xmath65_{{\rm dla}}= 0 $ ] . this method is conceptually similar to that originally used by pettini et al . ( 1994 ) for estimating the dust - to - gas ratio from cr and zn abundances . on the basis of what discussed in section 4.1 we have used the [ zn / h ] and [ cr / h ] ratios , and the [ zn / h ] and [ fe / h ] ratios , to derive @xmath50 from ( 19 ) for our sample of 17 dla systems . the results are shown in table 3 . in the 8 cases with available independent determinations we do not find a systematic difference between the results based on the [ zn / h ] and [ cr / h ] ratios and the results based on the [ zn / h ] and [ fe / h ] ratios . the average percent difference between the two cases is @xmath90 . by applying random error propagation to ( 19 ) , keeping fixed the galactic fractions in dust @xmath91 and @xmath92 , one finds about twice larger errors than in the case of the relation ( 18 ) . this is not surprising since two independent measurements are required to determine @xmath50 from ( 19 ) and only one to determine @xmath74 from ( 18 ) . systematically lower dust - to - gas ratios by about @xmath930.2 dex would be derived by assuming zinc completely undepleted . the resulting dust - to - gas ratios span a wide interval , from about 1% to 75% the galactic value , without a concentration around a particular value . the logarithmic average yields @xmath94 , which implies that the most probable interval lies between 2% and 25% the galactic value . this result is in good agreement with the most probable range found by pei , fall & bechtold ( 1991 ) in their study of qsos with and without foreground dla absorbers , which is between 5% and 20% . the combined sample used by these authors includes 22 qsos with foreground dla absorber , 7 of these systems being in common with the present sample . the quantitative agreement between the two independent extinction estimates supports the validity of the assumption that dust is responsible for the apparent deviations of the [ zn / cr ] and [ zn / fe ] ratios from solar composition . by assuming zinc completely undepleted , our results indicate a most probable range from 1% up to 16% of the galactic value . even if this interval still overlaps with the best range of pei , fall & bechtold ( 1991 ) , the mild zinc depletion adopted here yields a better agreement . once we have determined the dust - to - gas ratio , the metallicity corrected by dust can be determined from the relation @xmath95_{{\rm obs } } } + { \tilde{k}}\cdot f_{{{\rm y}},{{\rm ism}}}\ ] ] which results by applying eq . ( 12 ) to the element used as the reference for the metallicity level . the dust - corrected metallicities of our sample are shown in table 3 . since fe , cr and zn trace each other , equivalent results are obtained by using any of the three elements in eq . ( 20 ) . in practice we have used the [ zn / h ] ratios for which the correction term is smaller . on the average the corrected metallicities differ by only @xmath72 + 0.12 dex from the observed [ zn / h ] ratios , which would represent the real metallicity if zinc were completely undepleted . owing to the small amount of the correction , the use of the dust corrected metallicities of table 3 does not affect significantly the values of metallicities derived for dla absorbers . it is worth noting , however , that even a small amount of zinc depletion in the galaxy still implies a small but measurable amount of depletion in dla absorbers . this can be understood from our previous discussion in section 3 . an important by - product of the present investigation is the determination of dust - to - gas ratios and metallicities in individual dla systems . we start by analysing the dust - to - metals ratio @xmath96 , which is less affected by error propagation than the dust - to - gas ratio @xmath50 . the logarithmic average yields @xmath97 , which implies a typical value around 62% and the most probable range between 42% and 89% the galactic dust - to - metals content . an attempt to find correlations of @xmath74 with metallicity or with redshift gives negative results . we conclude that in the present sample of 17 dla systems _ the dust - to - metals ratio is approximately constant and lower than in the galaxy . _ the average dust - to - metals ratio that we find is comparable with the value @xmath98 obtained by pettini et al . ( 1997 ) by assuming zinc completely undepleted . the constancy of the @xmath96 ratio implies that dust - to - gas ratio and metallicities are correlated in dla systems . this can be seen in fig . 2 , where we plot @xmath50 versus metallicity @xmath51 for our sample of 17 dla systems . filled diamonds represent the 8 systems with most accurate measurements , based on two independent @xmath50 determinations in each case . the other systems , with one single @xmath50 determination are indicated with empty diamonds . the linear regression through the 17 dla data points confirms the existence of a correlation , with slope 1.16 and correlation coefficient @xmath99 ( dotted line in the figure ) . by using only the 8 systems with most accurate measurements we find a correlation with slope 0.98 and @xmath99 ( dashed line ) . the slope is therefore 1 within the errors , as expected for a constant @xmath96 ratio . the fact the dust content follows the metal content gives fresh support to the idea that dla systems are originated in galaxies : dust would be produced in the diffuse medium following the injection of metals from the stellar component . the constancy of the @xmath96 ratio has important implications in the framework of the model of cosmic chemical evolution developed by pei & fall ( 1995 ) . in fact , in that model _ it is assumed _ that the dust - to - metals ratio is constant in dla absorbers . our study _ shows _ that this basic assumption of the model is essentially correct . the range of redshifts investigated here is probably insufficient to detect variations of the dust - to - metals content that may be expected in the course of the cosmic evolution . it is worth noting that the constancy of the dust - to - metals ratio could have been predicted from the remarkably small scatter of the abundances measured in different dla systems . in fact , each one of the ratios [ si / fe ] , [ cr / fe ] and [ ni / fe ] from the compilation by lu et al . ( 1996 ) has dispersion @xmath100 0.2 dex , the most remarkable case being [ si / fe ] with @xmath101 = 0.11 dex in 12 dla systems . the compilation of [ zn / cr ] data by pettini et al . ( 1997 ) gives @xmath101 = 0.18 dex for 14 systems . within the typical measurement error the observed abundances in the gas phase are essentially constant this remarkable result can not be explained only by assuming that dla systems have equal intrinsic abundances and equal dust grains properties . even in this case the fractions in dust , and therefore also the observed abundances , are expected to depend on the dust - to - metals ratio , as can be seen from eq . ( 9 ) . since metallicity and dust content can significantly vary among dla systems , the only way to explain the constancy of the observed abundances is that @xmath102 is roughly constant in different absorbers , as we indeed find . taken at face values , the positive [ si / fe ] and [ s / fe ] ratios and the negative [ mn / fe ] ratios observed in dla systems suggest an abundance pattern similar to that found in the galactic halo ( lu et al . 1996 ) . however , this interpretation requires that dust is completely absent from the absorbers and is unable to explain why the iron - peak ratios [ zn / fe ] and [ ni / fe ] are different from zero . this inconsistency , together with the agreement between direct and indirect estimates of the dust - to - gas ratios discussed in section 4.3 , suggests that any interpretation of the observed abundances must take into account the presence of dust . in principle , the dust in dla absorbers could differ from the dust in the galaxy and an overall halo - like abundance pattern could be hidden by the peculiar properties of the dust . ( 9 ) shows that one can obtain a halo - like pattern from the observed quantities by playing with the chemical composition of the dust . nevertheless , since the condition @xmath103 must be satisfied , one will be also obliged to play with other parameters , such as @xmath24 , in order to obtain meaningful results ( it is possible to see that this is in fact the case for mn ) . therefore , only by invoking a set of ad hoc requirements on the chemical composition and on the extinction properties of the dust grains is it possible to argue that the intrinsic abundances follow a halo - like pattern . by contrast , _ the results shown in fig . 1 have been derived without any fine tuning of the input parameters _ , i.e. the galactic fractions in dust , _ which are known from galactic interstellar studies_. our results indicate that all the dla systems for which good quality abundance determinations are available show an elemental abundance pattern approximately solar , within at most @xmath2 dex ( see however the comments on nitrogen in section 4.2 ) . the absorbers included in this sample span a wide range of metallicities ( @xmath104 ) and redshifts ( @xmath105 ) . even if table 4 includes only 5 [ si / fe ] determinations , the remarkably similar [ si / fe ] values reported by lu et al . ( 1996 ) in other systems without zn measurement , together with the approximately constant dust - to - metals ratio ( section 5 ) suggest that the lack of [ @xmath0/fe ] enhancement is common to all the 12 absorbers with observed [ si / fe ] ratio . similarly , the mild [ mn / fe ] underabundance ( @xmath106 dex ) is probably present in all the 10 systems for which measurements are available and not only in the 2 systems shown in table 4 . future measurements for a statistically significant number of systems are required to see how general the solar - like ratios are . as mentioned in section 4.2 , nitrogen observations do show deviations from solar relative abundances . in addition , dla systems with intrinsically enhanced [ @xmath0/fe ] ratios might be hidden by selection effects . for instance , damped absorbers with high dust content could systematically obscure the background qso ( pei & fall 1995 ) and this fact could affect the probability of detecting cases with enhanced [ @xmath0/fe ] ratios . similarities between properties observed in dla absorbers and in gas rich dwarf galaxies , including the presence of large amounts of gas , the low level of metallicity , and the complexity of the line profiles had already been pointed out by york et al . pettini , boksenberg & hunstead ( 1990 ) noted , in addition , a similarity in the star formation rate and suggested that dla absorbers might arise predominantly in dwarf galaxies whose properties have not changed significantly from @xmath107 to the present epoch . according to tyson ( 1988 ) the frequency of detection of dla absorbers can be naturally explained by dwarf galaxies with no need to invoke extremely large hi diameters for the progenitors of present - day spirals . recent studies of the luminosity function of galaxies based on deep field observations ( campos 1997 ; zucca et al . 1997 ) indicate a steepening of the faint end of the luminosity function consistent with that suggested by tyson ( 1988 ) . in addition to the above arguments we note here that ( i ) the general lack of detection of molecules in dla absorbers ( see levshakov et al . 1992 and refs . therein ) is consistent with the low molecular content apparently typical of dwarf galaxies ( israel , tacconi & baas 1995 ) ; ( ii ) the lack of the extinction bump at @xmath72 2175 in dla absorbers ( pei , fall , & bechtold 1991 ) is characteristic of the extinction curves of the magellanic clouds . finally , we note that the high fraction of metal - poor stars predicted for the descendant of dla absorbers by lanzetta , wolfe & turnshek ( 1995 ) is consistent with most dla absorbers being metal - poor dwarf galaxies . the high fraction of metal - poor stars predicted by the model of lanzetta , wolfe & turnshek ( 1995 ) is not seen in the milky way and this inconsistency is known as the `` cosmic g - dwarf problem '' . to solve this problem it has been suggested that a large fraction of the gas in dla absorbers is consumed in forming stars in galactic bulges and spheroids rather than in galactic disks . however , this explanation conflicts with the low mass fraction of the halo and of the bulge ( bahcall & soneira 1980 ; rich 1992 ) relative to the total mass of the milky way , and is not supported by the range of abundances recently found in the galactic bulge ( rich 1992 ) . if dla absorbers are predominantly dwarf galaxies , then there is no `` cosmic g - dwarf problem '' because the stars of the descendant of dla absorbers do have low metallicities . we discuss now the observational evidences concerning the nature of dla galaxies related to the results of the present work . the present results suggest that dla galaxies commonly show a nearly solar [ @xmath0/fe ] ratio at a low level of metallicity . this property _ does not show evolution _ over a wide range of metallicities ( @xmath104 ) and redshifts ( @xmath105 ) . the lack of evolution up to high levels of metallicity might suggest that the population of absorbers is essentially unevolved to the present time . this possibility , however , can only be investigated by studying a sample of low redshift dla absorbers . the nearly solar [ @xmath0/fe ] ratios do not support an origin _ in spirals like the milky way , where the enhancement of the _ [ @xmath0/fe ] _ ratio is found not only in halo stars _ ( ryan et al . 1996 ) , _ but also in metal - poor stars of the disk _ ( edvardsson et al . it is unlikely therefore that dla absorbers arise in massive protogalactic disks . galaxies with low metallicities and without [ @xmath0/fe ] enhancement are known to exist , as demonstrated for instance by abundance studies of the magellanic clouds ( wheeler , sneden & truran 1989 ; p. 315 ) . models of chemical evolution developed to explain the elemental abundances observed in blue compact and dwarf irregular galaxies show that it is possible to obtain low [ @xmath0/fe ] ratios at low metallicities when many episodes of star formation and differential galactic winds are assumed ( marconi , matteucci & tosi 1994 ) . similar models have been applied to explain the abundances observed in some dla systems ( matteucci , molaro & vladilo 1997 ) . from a recent analysis pettini et al . ( 1997b ) have concluded that the metallicity distribution of dla absorbers is different from that found in the stellar populations of the milky way . as we mentioned in section 4.4 , the dust - corrected metallicities derived here are essentially equal to the direct [ zn / h ] measurements and therefore the conclusions of pettini et al . ( 1997b ) are unchanged . the dust - to - metals ratios of dla absorbers are systematically lower than in the milky way ( section 4.1 ) , even at metallicities comparable to that attained by the galactic disk . similarly low dust - to - metals ratios are instead found in the magellanic clouds . this can be seen by combining the typical metallicities of the lmc and smc , [ fe / h ] = @xmath108 and @xmath109 , respectively ( wheeler , sneden & truran 1989 ; p. 315 ) , with their typical gas - to - dust ratios , @xmath20/e(b - v ) @xmath110 and @xmath111 @xmath4 mag@xmath112 , respectively ( lequex 1989 ) . these last ratios yield @xmath113 0.24 and 0.096 , respectively , once normalized to the standard galactic value @xmath20/e(b - v ) @xmath114 @xmath4 mag@xmath112 . the magellanic ratios follow remarkably well the behaviour typical of dla systems , as we show in fig . 2 , and this is consistent again with an origin of dla absorbers in dwarf galaxies . this result is at variance with the claim by roth & blades ( 1997 ) that the magellanic dust - to - metals ratios are closer to milky way rather than to dla values . the claim is based on three measurements of the magellanic [ cr / zn ] ratio , two of which are in fact consistent , within the error bars , with dla values . the third measurement , towards sk @xmath11573 , gives a [ cr / zn ] ratio lower than in dla systems , but in this case the magellanic znii and crii column densities are exceptionally high , not only compared to the other two magellanic lines of sight , but even compared to the 20 galactic lines of sight observed with the hst ghrs ( roth & blades 1995 ) . the lmc path toward sk @xmath11573 is clearly atypical and might hide an exceptional amount of dust . the few available magellanic [ cr / zn ] measurements , therefore , do not contradict our conclusion that magellanic dust - to - metals ratios are comparable to dla ratios . a direct answer on the morphology of dla galaxies can be obtained , in principle , from imaging observations of the close environment of qsos showing dla absorption . a study of this kind has been performed by le brun et al . ( 1997 ) for 8 absorbers at @xmath5 in 7 fields observed with the hst / wfpc2 . the candidate absorbing galaxies display a variety of morphological types , including 3 compact galaxies , 2 low surface brightness ( lsb ) , and 3 spirals . our sample of table 3 includes four of the absorbers studied by le brun et al . ( 1997 ) : three are compact galaxies and one is a lsb galaxy . the sample of dust - corrected elemental abundances ( table 4 ) includes two of these absorbers and both are compact galaxies . therefore , also the available imaging data are consistent with an origin in dwarf galaxies of the absorbers of the present sample . also the kinematics gives important clues for understanding the nature of dla galaxies . the velocity spread of the absorption components seen through a randomly oriented galaxy along the line of sight is in fact expected to be higher for a fastly rotating spiral than for a slowly rotating dwarf . nevertheless , there are evidences of dwarfs with velocity spreads higher than expected from a slow rotation . for instance , velocities spreads of @xmath1 100 km s@xmath112 are found both from stellar ( westerlund 1990 ) and interstellar ( blades et al . 1988 , songaila et al . 1986 ) studies of the magellanic clouds . the compact galaxies identified by le brun et al . ( 1997 ) as the @xmath116 absorber toward qso 0454 + 039 and the @xmath117 absorber toward qso 1331 + 170 have velocity spreads of 150 km s@xmath112 and 100 km s@xmath112 , respectively ( see lu et al . such relatively high velocity widths could be due to the presence of flows or systematic motions superposed to the low rotation pattern of dwarf galaxies ( see e.g. sahu & blades 1997 ) . six out of the nine absorbers in our sample of table 4 have velocity spreads of at most 100 - 150 km s@xmath112 and are therefore consistent with an origin in dwarf galaxies . the remaining three absorbers ( @xmath118 in qso 0201 + 363 , @xmath119 in qso 0528 - 250 , and @xmath120 in qso 2206 - 199 ) show instead velocities of 200 km s@xmath112 or even higher ( lu et al . 1996 , prochaska & wolfe 1996 , prochaska & wolfe 1997 ) . such a high velocity width is quite unlikely for a single dwarf galaxy , but is not a conclusive evidence for a fastly rotating disk , since multiple intervening galaxies at similar redshift can also produce a broad absorption profile . in fact , lu , sargent & barlow ( 1997 ) present circumstantial evidence that the large velocity width of the @xmath119 system toward qso 0528 - 250 can be due to the presence of more than one galaxy or sub - galactic fragment along the line of sight . in our sample we do not see systematic differences in the elemental abundance patterns between the cases with moderate and large velocity spread . it is possible that the few cases with high velocity spread are due to the superposition of low mass intervening galaxies . from a statistical analysis of sample of 17 dla absorbers prochaska & wolfe ( 1997b ) claim that models of slowly rotating dwarf galaxies ( @xmath121 km s@xmath112 ) are excluded at 97% confidence level , while models with rapidly rotating disks ( @xmath122 km s@xmath112 ) are consistent with the data at high confidence level . these authors do not exclude , however , that some of the cases with higher velocity spreads might arise from multiple intervening galaxies of low mass . an inspection of the velocity profiles of their sample reveals that the radial velocity spread is @xmath123 100 km s@xmath112 in 10 systems , between @xmath72 100 and @xmath72 150 km s@xmath112 in 2 systems , and @xmath124 200 km s@xmath112 in only 5 systems . in the present work the general equations that link elemental abundances observed in the gas phase with real abundances of dla absorbers ( gas plus dust ) have been presented . by assuming that dust is of galactic type , simplified equations have been obtained in which the observed abundances are expressed as a function of the dust - to - metals ratio , metallicity , element - to - element abundance pattern , and of galactic interstellar dust parameters , namely the fractions in dust of the elements of interest . the behaviour of the dust depletions in dla absorbers has been investigated for different sets of element - to - element abundances and of dust - to - metals ratios . the dust depletion pattern does not scale linearly in logarithm with the galactic pattern of depletion , contrary to what assumed in some previous works . instead , the differential depletion between pairs of elements tends to vanish as the dust - to - metals ratio decreases . in addition , inversions in the differential depletion pattern may appear if the dla abundance pattern is not solar . a method has been presented for determining dust - to - metals ratios and dust - to - gas ratios in individual dla absorbers from the abundances of elements with similar nucleosynthetic history . this method has been applied to a sample of 17 dla absorbers with available zn , cr an fe measurements , assuming that these iron - peak elements trace each other down to the metallicities of dla systems , as they do in metal - poor stars of the galaxy . the resulting dust - to - gas ratios are mostly distributed between 2% and 25% of the galactic value , in good agreement with the most probable range of dust - to - gas ratios obtained by pei , fall & bechtold ( 1991 ) from their reddening study of qsos with foreground dla absorption . dust - to - gas ratio and metallicity are found to be correlated in dla galaxies , with a typical dust - to - metals ratio @xmath72 60% of the galactic value . this suggests that the metals injected from the stellar component are systematically removed from gas to dust with an efficiency approximately constant but lower than in the milky way . the correlation between dust - to - gas ratio and metallicity is relevant to models of cosmic chemical evolution of the type investigated by pei & fall ( 1995 ) . the approximate constancy of the dust - to - metals ratio explains why the element - to - element abundances observed in dla systems are remarkably constant in spite of the variable amount of dust in the absorbers . the resulting dust - to - metals ratios have been used to correct from dust effects the element - to - element abundances measured in a sub - sample of 9 dla absorbers . the dust - corrected [ @xmath0/fe ] ratios do not show evidence of the enhancement characteristic of metal - poor galactic stars , but have instead solar values within @xmath2 dex . the [ mn / fe ] ratio shows a mild underabundance , less marked than that found in metal - poor galactic stars . the [ ni / fe ] ratio is essentially solar , with a weak evidence for a modest underabundance . these results rely on the assumption that dust in dla absorbers has same chemical composition as galactic interstellar dust . the available measurements of the [ s / zn ] ratio , a proxy of the [ @xmath0/fe ] ratio virtually unaffected by dust , support the validity of the dust correction applied here ( molaro , centurin & vladilo , 1997 ) . the abundance pattern of dla absorbers does not show evidence of evolution in the range of redshifts @xmath125 and of metallicities @xmath126 investigated . studies of low redshift absorbers are required to establish whether evolution is present or not closer to the present time ( vladilo et al . 1997a ) . the dust - corrected pattern of abundances indicates that the chemical history of a large fraction of dla galaxies is different from that experienced by the milky way halo or disk . the results of the present work are consistent with dla absorbers being predominantly dwarf rather than spiral galaxies . evidence in this sense comes from a wide spectrum of observational properties which include , in addition to the abundance pattern , the dust - to - metals ratio and the morphologycal type of the intervening galaxies at @xmath5 . the kinematics of the absorbers is in most cases consistent with an origin in dwarfs , even though an origin in spirals gives a more natural explanation of the highest observed velocity spreads . it is necessary to study a larger sample of dla systems by means of spectroscopic and imaging techniques in order to determine with statistical accuracy the fraction of spiral galaxies in this class of qso absorbers . at the same time , it is important to understand whether the low fraction of identified spirals is real or , instead , is due to some selection effect or to a peculiar dust composition that would alter the conclusions of the present work . if the fraction of spirals will be confirmed to be low , a mass of dla galaxies lower than expected for proto - spirals objects would help to reconcile cold plus hot dark matter ( chdm ) cosmology with observed properties of high redshift absorbers ( see klypin at al . 1995 and refs . therein ) . the interpretation of this work has benefited from important discussions with p. molaro . i wish to thank p. bonifacio and m. centurin for their critical reading of the manuscript . discussions with s. bardelli , s. borgani , f. matteucci , and s. ortolani have helped to clarify specific topics of this work . i also thank the referee , l. lu , for his constructive remarks and for anticipating some of his results in advance of publication . zucca e. , zamorani g. , vettolani g. , cappi a. , merighi r. , mignoli m. , stirpe g. m. , macgillivray h. , collins c. , balkowski c. , cayatte v. , maurogordato s. , proust d. , chincarini g. , guzzo l. , maccagni d. , scaramella r. , blanchard a. , ramella m. , 1997 , , in press ( astro - ph/9705096 ) . si & 0.63 & + 0.27 & -0.43 & -0.28 & -0.16 & -0.18 & -0.13 & -0.08 s & 0.12 & + 0.41 & -0.06 & -0.04 & -0.03 & -0.02 & -0.02 & -0.01 ti & 0.91 & + 0.22 & -1.02 & -0.49 & -0.26 & -0.35 & -0.23 & -0.14 mn & 0.88 & -0.35 & -0.92 & -0.47 & -0.25 & & & -1.82cr & 0.92 & -0.07 & -1.10 & -0.51 & -0.27 & & -0.72 & -0.34 fe & 0.94 & & -1.22 & -0.53 & -0.28 & -1.22 & -0.53 & -0.28 ni & 0.97 & -0.02 & -1.46 & -0.56 & -0.29 & & -0.62 & -0.31 zn & 0.35 & + 0.04 & -0.19 & -0.13 & -0.08 & -0.17 & -0.12 & -0.08 lccccccccccc 0056 + 014 & 2.777 & -1.23 & 1 & -1.35 & 1 & & & & & & 0100 + 130 & 2.309 & -1.55 & 2 & -1.79 & 2 & -1.95 & 3 & -2.13 & 2 & & 0112 + 029 & 2.422 & -1.15 & 4,1 & -1.65 & 4,1 & & & & & & 0201 + 363 & 2.462 & -0.26 & 5 & -0.90 & 5 & -0.80 & 5 & -1.00 & & & 0302 -223 & 1.009 & -0.51 & 6 & -0.90 & 6 & & & & & & 0454 + 039 & 0.859 & -0.83 & 7,1 & -1.02 & 7,1 & -0.97 & 8 & & & -1.22 & 8 0458 -020 & 2.039 & -1.23 & 4,1 & -1.59 & 4,1 & & & -1.90 & 9 & & 0528 -250 & 2.811 & -0.76 & 8 & -1.23 & 8 & -1.23 & 8 & -1.56 & 8 & & 0841 + 129 & 2.374 & -1.35 & 1 & -1.64 & 1 & & & & & & 0935 + 417 & 1.372 & -0.80 & 10,1 & -0.90 & 10,1 & -0.96 & 10 & & & -1.24 & 10 1104 -180 & 1.661 & -0.80 & 11,1 & -1.29 & 11,1 & & & & & & 1151 + 068 & 1.773 & -1.56 & 1 & -1.64 & 1 & & & & & & 1223 + 175 & 2.465 & -1.68 & 4,1 & -1.82 & 4,1 & & & & & & 1328 + 307 & 0.692 & -1.21 & 12,1 & -1.66 & 12,1 & -1.80 & & & & & 1331 + 170 & 1.776 & -1.27 & 8 & -2.10 & 8 & -2.03 & 8 & -2.20 & & & 2206 -199 & 1.920 & -0.38 & 13 & -0.58 & 13 & -0.68 & 13 & -1.01 & 13 & & 2231 -001 & 2.066 & -0.88 & 8 & & & -1.14 & 8 & -1.53 & 8 & & 0056 + 014 & 2.777 & 0.26 & 0.37 & & & 0.025 & & -1.60 & -1.17 0100 + 130 & 2.309 & 0.46 & 0.59 & 0.64 & 0.75 & 0.021 & 0.029 & -1.60 & -1.43 0112 + 029 & 2.422 & 0.74 & 0.85 & & & 0.085 & & -1.07 & -1.00 0201 + 363 & 2.462 & 0.84 & 0.92 & 0.76 & 0.85 & 0.747 & 0.667 & -0.15 & -0.10 0302 -223 & 1.009 & 0.64 & 0.76 & & & 0.323 & & -0.49 & -0.37 0454 + 039 & 0.859 & 0.38 & 0.51 & 0.29 & 0.40 & 0.092 & 0.070 & -1.09 & -0.75 0458 -020 & 2.039 & 0.61 & 0.74 & & & 0.059 & & -1.23 & -1.10 0528 -250 & 2.811 & 0.72 & 0.83 & 0.70 & 0.81 & 0.203 & 0.196 & -0.70 & -0.61 0841 + 129 & 2.374 & 0.53 & 0.65 & & & 0.038 & & -1.42 & -1.23 0935 + 417 & 1.372 & 0.22 & 0.32 & 0.33 & 0.44 & 0.057 & 0.083 & -1.13 & -0.73 1104 -180 & 1.661 & 0.73 & 0.84 & & & 0.189 & & -0.72 & -0.65 1151 + 068 & 1.773 & 0.18 & 0.27 & & & 0.008 & & -2.09 & -1.52 1223 + 175 & 2.465 & 0.30 & 0.41 & & & 0.010 & & -2.00 & -1.61 1328 + 307 & 0.692 & 0.70 & 0.81 & 0.79 & 0.88 & 0.070 & 0.078 & -1.13 & -1.06 1331 + 170 & 1.776 & 0.93 & 0.98 & 0.88 & 0.94 & 0.081 & 0.076 & -1.11 & -1.09 2206 -199 & 1.920 & 0.40 & 0.53 & 0.53 & 0.65 & 0.271 & 0.354 & -0.50 & -0.28 2231 -001 & 2.066 & & & 0.48 & 0.60 & & 0.101 & -0.99 & -0.78 0100 + 130 & 2.309 & & & + 0.41 & 3 & & & & & -0.18&2,3 & & & & + 0.01 & & & & & & -0.05 & 0201 + 363 & 2.462 & + 0.42 & 5 & & & & & & & -0.20 & 5 & & 0.00 & & & & & & & & -0.02 & 0454 + 039 & 0.859 & & & & & & & -0.25 & 8 & & & & & & & & & & -0.14 & & & 0458 -020 & 2.039 & & & & & & & & & ( -0.67)&1,9 & & & & & & & & & & -0.06 & 0528 -250 & 2.811 & + 0.48 & 8 & + 0.35 & 8 & & & & & -0.33 & 8 & & + 0.08 & & -0.20 & & & & & & -0.05 & 0935 + 417 & 1.372 & & & & & + 0.23 & 10 & -0.28 & 10 & & & & & & & & + 0.15 & & -0.17 & & & 1331 + 170 & 1.776 & + 0.18 & 8 & + 0.76&8,14 & & & & & -0.17 & 8 & & -0.13 & & -0.16 & & & & & & 0.00 & 2206 -199 & 1.920 & + 0.27 & 13 & & & -0.10 & 13 & & & -0.34 & 13 & & + 0.08 & & & & -0.05 & & & & -0.11 & 2231 -001 & 2.066 & + 0.26 & 8 & & & & & & & -0.39 & 8 & & + 0.07 & & & & & & & & -0.12 &	the effects of the dust on the determination of elemental abundances in damped ly @xmath0 ( dla ) absorbers are investigated . relations between the observed abundances measured in the gas phase and the overall abundances ( gas plus dust ) are derived as a function of dust - to - gas ratio , metallicity , element - to - element abundance pattern , average extinction coefficient of dust grains , and chemical composition of dust grains . a method is presented for determining dust - to - gas ratios , dust - to - metals ratios , and dust - corrected relative abundances in dla absorbers by assuming dust of galactic type and constant abundance ratios between iron - peak elements . the method is applied to a sample of 17 dla absorbers with available zn , cr and/or fe measurements . the resulting dust - to - gas ratios are mostly distributed between 2% and 25% of the galactic value , in good quantitative agreement with the results from reddening studies of qsos with foreground dla absorption . a correlation is found between dust - to - gas ratio and metallicity in dla galaxies , with a typical dust - to - metals ratio of @xmath1 60% the galactic value . the derived dust - to - metals ratios are then used to correct from the effects of dust the abundance ratios [ si / fe ] , [ s / fe ] , [ ti / fe ] , [ mn / fe ] , [ ni / fe ] available for a sub - sample of 9 absorbers . the [ @xmath0/fe ] ratios corrected from dust do not show the enhancement characteristic of metal - poor galactic stars , but instead have essentially solar values , within @xmath2 dex . this suggests that the chemical history of dla absorbers is different from that experienced by the milky way . evidences that point to dwarf galaxies , rather than to spiral galaxies , as important contributors to the dla phenomenon are summarized .
hopf soliton solutions arise as topological solitons in the skyrme - faddeev model @xcite a nonlinear @xmath0 sigma model in @xmath1-dimensional space - time whose lagrangian is modified by an additional term quartic in its field derivatives . extensive numerical simulations @xcite of the highly nonlinear classical field equations have revealed a very rich spectrum of solutions that are classified by their integer - valued hopf charge . for hopf charges up to 16 a variety of static , stable minimum - energy solutions with the structure of closed strings , twisted tori , linked loops , and knots have been identified . these stringlike solitons might be candidates to model glueball configurations @xcite in qcd or may arise in two - component bose condensates @xcite . in this paper we investigate the effect of isospin on classical hopf soliton solutions . in analogy to the conventional @xmath2 skyrme model we use the collective coordinate method @xcite to construct hopf solitons of well - defined , nonzero isospin : we parametrize the isorotational zero modes of a hopf configuration by collective coordinates , which are then taken to be time dependent . this gives rise to additional dynamical terms in the hamiltonian , which can then be quantized following semiclassical quantization rules . a simplification that is often made in the literature @xcite is to apply a simple adiabatic approximation to the ( iso)rotational zero modes of the soliton by assuming that the soliton s shape is rotational frequency independent . the limitations of this rigid body approach were pointed out by several authors @xcite . in this paper we perform numerical computations of isospinning hopf solitons with hopf charges up to 8 in the full three - dimensional classical field theory without applying the rigid body approximation and without imposing symmetry constraints on the isospinning hopf configurations . it turns out that the skyrme - faddeev model with its rich topology of minimum - energy solutions , often of comparable energy , allows for `` transmutations '' when isospin is added and even for the formation of new , metastable hopf solutions . this paper is organized as follows . in sec . [ sec_sf ] we briefly review the skyrme - faddeev model and describe how hopf solitons acquire isospin within the collective coordinate approach . then , in sec . [ sec_ic ] we set up appropriate initial conditions which are used in sec . [ sec_stat ] to compute hopf configurations of zero isospin . the effect of isospin on these hopf soliton solutions is studied in sec . [ sec_res ] . we conclude with sec . [ sec_con ] . the lagrangian density of the skyrme - faddeev model @xcite in @xmath1 dimensions takes in terms of the real three - component unit vector @xmath3 the form @xmath4 to stabilize isospinning hopf configurations we modified in ( [ lag_hopf_phi ] ) the usual skyrme - faddeev model by adding a mass term @xmath5 to the @xmath0 sigma model and skyrme term . here we will consider the following @xmath6 symmetry breaking potentials : @xmath7 where @xmath8 is a rescaled mass parameter . the potential @xmath9 has one vacuum for @xmath10 , whereas @xmath11 has two vacua : for @xmath10 and @xmath12 . the planar version of ( [ lag_hopf_phi ] ) with @xmath13 corresponds to the old baby skyrme model @xcite , and the one with @xmath14 reproduces the new baby skyrme model @xcite . the normalization in ( [ lag_pot ] ) is choosen so that for @xmath15 both potentials show the same asymptotic behaviour , explicitely given by @xmath16 . the lagrangian ( [ lag_hopf_phi ] ) admits topologically nontrivial , stringlike , finite - energy configurations due to the third homotopy group of the 2-sphere being nontrivial , @xmath17 . this can be seen as follows . a static finite - energy configuration requires the boundary condition @xmath18 as @xmath19 for all time @xmath20 . hence this boundary condition on the field @xmath21 defines a mapping @xmath22 , and the field configurations can be classified topologically by the homotopy group @xmath17 . the topological invariant associated with each static field configuration is known as the hopf charge @xmath23 . it can be interpreted geometrically as the linking number of two loops obtained as the preimages of any two generic distinct points on the target 2-sphere . the position curve of the soliton is defined as the set of points where the field is as far as possible from the boundary vacuum value @xmath24 . thus it is given by the preimage of the point @xmath25 , which is antipodal to the vacuum value . when we visualize hopf solitons position curves , we usually display for clarity tubelike isosurfaces with @xmath26 , where @xmath27 is chosen to be small . similarly the linking curve can be illustrated graphically by plotting an isosurface of the preimage of the vector @xmath28 . the overall factor @xmath29 in ( [ lag_hopf_phi ] ) is motivated by ward s conjecture @xcite that with the normalization ( [ lag_hopf_phi ] ) the vakulenko - kapitanski lower bound @xcite on the energy @xmath30 of a hopf configuration with charge @xmath23 is given by @xmath31 the topological bound ( [ ward_bound ] ) has been shown to be compatible with fully three - dimensional numerical simulations carried out in the massless skyrme - faddeev model @xcite and in the massive one @xcite with potential @xmath32 included . the skyrme - faddeev model ( [ lag_hopf_phi ] ) can be expressed in analogy to the conventional @xmath2 skyrme model @xcite in terms of the @xmath2-valued hermitian scalar field @xmath33 . the ansatz for the dynamical soliton field adopted in the collective coordinate quantization @xcite is given by @xmath34 where we have promoted the collective coordinate @xmath35 to a time - dependent dynamical variable and ignored the translational and rotational degrees of freedom . @xmath36 describes the isorotational fluctuations about the classical minimum - energy solution @xmath37 . substituting ( [ u_dyn ] ) in ( [ lag_hopf_phi ] ) and defining the body - fixed angular velocities via @xmath38 the skyrme - faddeev lagrangian takes the form @xmath39 where the hopf soliton mass @xmath30 is given by @xmath40+v\left(\boldsymbol{\phi}\right)\bigg\}\,\text{d}^3x\ , , \label{hopf_mass_phi}\end{aligned}\ ] ] and the moment of inertia tensors is @xmath41 [ hopf_inertia_phi ] the momentum conjugate to @xmath42 is the body - fixed isorotation angular momentum @xmath43 defined via @xmath44 in this article , we choose the @xmath45 axis as our rotation axis . using gradient - based mehods we search for hopf configurations @xmath21 of a given topological charge @xmath23 , which minimize @xmath46 where the rotation frequency @xmath47 is calculated at each time step for a fixed @xmath48 as follows @xmath49 we create suitable initial field configurations with nontrivial hopf charge @xmath23 by using the approach presented in ref . the basic idea is to approximate the hopf configuration by rational maps @xmath50 , that is , a mapping from the three - sphere to the complex projective line . this approach enables us to set up initial conditions for knotted , linked and axial hopf configurations with energies reasonably close to the suspected minimum energy solutions . these initial conditions can then be relaxed using a modified version of the energy minimization algorithm @xcite originally designed to study skyrmion solutions . first we compactify @xmath51 to a unit 3-sphere @xmath52 via a degree one spherically equivariant map given by @xmath53 where @xmath54 , @xmath55 and @xmath56 are complex coordinates on the unit 3-sphere ( with @xmath57 ) . here @xmath58 is a monotonically decreasing profile function with boundary conditions @xmath59 and @xmath60 . in our simulations we use a simple linear profile function @xmath61 for @xmath62 and @xmath63 for @xmath64 with @xmath65 . approximate hopf solutions can be obtained by writing the stereographic projection of the field @xmath21 @xmath66 as a rational function of the complex variables @xmath67 and @xmath68 @xmath69 where @xmath70 and @xmath71 are polynomials in @xmath67 and @xmath68 . there are three different solution types which will be used as initial field configurations for our energy relaxation simulations : * toroidal fields of solution type @xmath72 can be obtained by setting @xmath73 where @xmath74 . the integer pair @xmath75 counts the angular windings around the two cycles of the torus . an axially symmetric hopf configuration of the type @xmath72 can be described @xcite by a baby skyrmion solution with winding number @xmath76 , which is embedded in the @xmath1-dimensional skyrme - faddeev model ( [ lag_hopf_phi ] ) along a closed curve and with its internal phase rotated through an angle @xmath77 as it travels around the circle once . the hopf charge @xmath23 associated with such an unlinked hopf configuration ( [ axial_rational_hopf ] ) is given by @xmath78 . * @xmath79-torus knots @xmath80 are described by the mapping @xmath81 where @xmath82 is a positive integer , @xmath83 is a non - negative integer and @xmath84 are coprime positive integers with @xmath85 . the rational map ( [ knot_rational_hopf ] ) generates a knot lying on the surface of an unknotted torus and winding @xmath86 and @xmath87 times about the torus circumferences . fields of type @xmath88 have topological charge @xmath89 @xcite . * linked hopf initial configurations of the type @xmath90 can be constructed , when the denominator of ( [ knot_rational_hopf ] ) is reducible . following the notation of @xcite , @xmath70 and @xmath71 label the charges of the two disconnected components that form the link , and the additional linking number of each component due to its linking with the other is denoted by the superscripts @xmath82 and @xmath83 . the total hopf charge of a field @xmath90 is @xmath91 . in particular , in this paper we will use the rational map @xmath92 to produce smooth initial linked configurations of solution type @xmath93 and hopf charge @xmath94 . in the following section , we compute minimum - energy hopf solutions for potential @xmath9 and @xmath11 using a relaxation algorithm with initial conditions constructed from the rational maps ( [ axial_rational_hopf ] ) , ( [ knot_rational_hopf ] ) and from linked configurations like e.g. , ( [ link_rational_hopf ] ) . to avoid saddle point solutions of the skyrme - faddeev energy functional @xmath30 , we explicitely add , in a similar way to @xcite , symmetry - breaking , nonaxial perturbations to our initial conditions . to find the stationary points of the energy functional @xmath30 , we solve the associated euler - lagrange equations numerically . the field equations can be implemented analogous to ref . @xcite @xmath95 where @xmath96 is a symmetric matrix . the dissipation @xmath97 in ( [ num_flow ] ) is added to speed up the relaxation process , and the lagrange multiplier @xmath98 imposes the unit vector constraint @xmath99 . we do not present the full field equations here since they are cumbersome and not particularly enlightening . the initial configuration is then evolved according to the flow equations ( [ num_flow ] ) . kinetic energy is removed periodically by setting @xmath100 at all grid points . all the simulations presented in the following use fourth order spatial differences on grids with @xmath101 points , a spatial grid spacing @xmath102 , and time step size @xmath103 . the dissipation is set to @xmath104 , and we choose the rescaled mass parameter @xmath105 throughout this paper . a summary of our relaxed configurations is given in table [ tab_hopf_en ] . each initial configuration is listed together with the final hopf configuration it evolves to . in fig . [ fig_hopf_link ] we display the linking structure of the minimum - energy configurations of charge @xmath106 obtained for potential @xmath9 . here , we visualize the field configurations by plotting isosurfaces of the points @xmath107 and @xmath108 with @xmath109 . our calculations with potential @xmath11 produce the same hopf solution types as for potential @xmath9 , the main difference being that the solitons are more compact . the minimal energy solutions of both massive models are very similar to the massless ones @xcite . .all initial conditions and final @xmath105 hopf configurations with their respective energies . @xmath110 and @xmath111 denote the soliton energy with potential @xmath9 and @xmath11 included , respectively . the superscript `` @xmath112 '' indicates that we applied nonaxial perturbations to the initial configuration . @xmath113 these configurations correspond to global energy minima for given hopf charge @xmath23 . recall that energies are given in units of @xmath29 . [ cols="^,^,^,^,^,^,^,^",options="header " , ] [ tab_hopf_en ] + + relaxing ( [ axial_rational_hopf ] ) with @xmath114 reproduces the @xmath115 static hopf configuration , which has for @xmath9 an energy @xmath116 and a moment of inertia @xmath117=0.500 , and this agrees well with @xmath118 stated in ref . @xcite . for comparison , substituting a spherically symmetric hedgehog form @xmath119 in ( [ hopf_mass_phi ] ) and minimzing the energy with respect to the profile function @xmath120 gives for the 1-hopf soliton solution an energy @xmath121 and a moment of inertia @xmath122 . the minimal energy @xmath123 hopf solitons are of the type @xmath124 axially symmetric configurations with the linking curve twisted two times around the position curve . applying nonaxial perturbations to an @xmath125 or a @xmath126 initial configuration , we find the @xmath127 3-hopf soliton solution to be of lowest energy for both potential choices . here , the tilde indicates that the position curve is not lying completely in the plane but it is bent . for completeness , we also include in tab . [ tab_hopf_en ] and in fig . [ fig_hopf_link ] minimal - energy configurations of solution type @xmath128 . these axial solutions are known to be unstable for @xmath129 @xcite . taking perturbed axially symmetric @xmath130 and knotted @xmath131 configurations as our initial conditions , we identify the bent axial solution @xmath132 as the global energy minimimum for @xmath133 and potential @xmath9 @xcite . the charge-4 configuration @xmath134 ( created from axial and linked initial conditions ) and @xmath130 are local energy minima . however , for potential @xmath11 the minima swap with @xmath134 becomes the lowest minimal - energy charge-4 soliton solution . for @xmath135 the minimal configuration in both massive models is a link of type @xmath136 , which we obtained by relaxing a perturbed trefoil knot @xmath137 . the charge-5 bent solution @xmath138 and the toroidal @xmath139 seem to be metastable local minima . for @xmath140 we find using a variety of initial conditions that the @xmath141 configuration has minimal energy , whereas the links @xmath142 , the bent unknot @xmath143 , and the rotationally symmetric unknot @xmath144 are only local minima @xcite . this differs from the massless skyrme - faddeev model where the link @xmath145 is the minimal - energy charge-6 soliton . similar to the massless case , the trefoil knot @xmath137 turns out to be the global minimum for @xmath146 in the massive models . charge-7 hopf solutions like the @xmath147 knot and the bent unknot @xmath148 represent local minima . finally , for @xmath149 we identify @xmath150 as the minimal - energy solution . for potential @xmath9 the trefoil knot @xmath151 can be seen within the numerical accuracy as an almost energy - degenerate state . the link @xmath152 which is the minimal - energy solution type in the massless model relaxes to @xmath150 . in fig . [ fig_ebound ] we show in analogy to ref . @xcite the normalized minimum energies @xmath153 for both potential choices . in both cases the energies of the ground - state hopf configurations ( filled circles ) follow @xmath154 . as already pointed out in ref . @xcite for the massless case , the energies for the @xmath155 configurations in the massive models are particularly low compared to the standard level . we verify in fig . [ fig_ebound ] that the normalized energies of the bent configurations @xmath156 with @xmath157 are well described by the linear @xcite fits @xmath158 and @xmath159 for @xmath9 and @xmath11 , respectively . a very similar fit ( @xmath160 ) is given in ref . @xcite for the bent unknots in the massless skyrme - faddeev model . for the planar configurations @xmath128 with @xmath161 we obtain @xmath162 for @xmath9 and @xmath163 for @xmath11 and @xmath164 . the corresponding quadratic fit for massless rotationally symmetric unknots is given as @xmath165 in ref . in this section , we present the results of our energy minimization simulations of isospinning hopf solitons with charges @xmath23 up to 8 . the variational equations derived from ( [ e_min ] ) are implemented in analogy to ( [ num_flow ] ) , where we include in @xmath166 the isorotational extra terms . we use the configurations obtained in the previous sections as our start configurations for vanishing angular momentum ( @xmath167 ) and increase @xmath168 in a stepwise manner . all simulation parameters are chosen as stated in sec . [ sec_stat ] . in particular , we use the mass parameter @xmath105 and work on grids containing @xmath101 lattice points with a lattice spacing @xmath102 . if not stated otherwise , we use @xmath13 as our potential term in ( [ lag_hopf_phi ] ) . note that for @xmath169 there exists a maximal frequency @xmath170 beyond which no stable isospinning hopf soliton solution exists . this upper limit follows from the stability analysis of the linearized euler - lagrange equations derived from ( [ e_min ] ) . we show in fig . [ ew_isospin_n1n3 ] the total energy @xmath172 as a function of the rotation frequency @xmath173 and the angular momentum @xmath168 for isospinning hopf solitons ( of type @xmath174 ) with charges up to 3 . the corresponding plots for the moment of inertia @xmath117 as function of @xmath173 are also presented . for all these configurations the solution type of the isospinning soliton is the same as the one in the static case , only the soliton s size grows with @xmath173 and @xmath168 . as expected , the energies and moment of inertia diverge for @xmath175 . + + * @xmath133 : the energy and moment of inertia plots for isospinning 4-hopf solitons ( @xmath177 ) are shown in fig [ ew_isospin_n4 ] . the @xmath178 configuration is found to be the solution type of lowest energy for all @xmath173 and @xmath168 . the @xmath179 soliton deforms for @xmath180 ( @xmath181 ) into a @xmath182 link , which means into a solution type which does not represent a local minimum in the static case ( @xmath183 ) . the isosurface plots in fig . [ iso_isospin_a22 ] illustrate the formation of the linked configuration as @xmath168 increases . * @xmath135 : we show in fig . [ ew_isospin_n5 ] the total energy @xmath172 of isospinning charge-5 hopf solitons ( @xmath184 ) as a function of the rotation frequency @xmath173 and the angular momentum @xmath168 . we observe that the energy curve @xmath185 of the linked unknot @xmath186 crosses the one of the bent ring @xmath138 at @xmath187 . for @xmath188 the bent ring becomes the new ground state for hopf charge @xmath135 . however , for fixed @xmath168 the linked configuration continues to be the lowest energy state , see fig . [ ew_isospin_n5 ] . * @xmath140 : our simulations of isospinning 6-hopf solitons ( @xmath189 ) are summarized by the energy curves in fig . [ ew_isospin_n6 ] . here , we see an example of transmutation : the @xmath190 configuration that is the ground state at @xmath183 transforms into a @xmath191 link when k increases . for @xmath192 ( @xmath193 ) the @xmath190 soliton has completely deformed into the link configuration that forms the new lowest energy state . the deformation process is visualized by the isosurface plots in fig . [ n6_def ] . bent hopf configurations of solution type @xmath194 and links of type @xmath195 have higher energies for all @xmath173 and @xmath168 . the linking curves in fig . [ n6_links_def ] show that the @xmath195 configuration is of the same qualitative shape for all @xmath173 and @xmath168 . * @xmath146 : we do not observe any crossing of the energy curves of isospinning @xmath196 and @xmath197 knot solutions . we find the @xmath196 knot as the state of lowest energy for all @xmath173 and @xmath168 . * @xmath149 : we display in fig . [ ew_isospin_n8 ] the total energies of isospinning @xmath198 and @xmath151 configurations as a function of angular frequency and momentum . at @xmath183 the configurations can be seen as energy - degenerate , but in the isospinning case we find that the @xmath151 solution type has a higher energy than the @xmath198 configuration . in fact , we can see that the @xmath198 solution slowly deforms into the knotted solution , as illustrated in fig . [ n8_def ] by plotting the linking structure . the transition occurs at @xmath199 ( @xmath200 ) . we have performed full three - dimensional numerical relaxations of isospinning soliton solutions in the skyrme - faddeev model with mass terms included . our computations of charge-4 , -6 and -8 solitons show that the qualitative shapes of internally rotating hopf solitons can differ from the static ( @xmath183 ) solitons . however , in most cases ( for hopf charges @xmath201 ) the solution types present at @xmath183 also exist for non - zero @xmath173 . the qualitative shape of the lowest energy configuration can be frequency dependent . the energy curves @xmath185 for a given @xmath23 can cross and minima can swap ( e.g. , @xmath135 ) . in summary , we distinguish three different types of behavior : * _ crossings of @xmath202 : _ the energy curves @xmath202 of hopf solitons for different solution types of the same charge @xmath23 can cross , which results in a rearrangement of the spectrum of minimal - energy configurations . our simulations on isospinning charge-5 solitons illustrate this : at @xmath183 the link @xmath186 is the lowest energy solution , but for @xmath187 its energy curve crosses that of the bent unknot @xmath203 . for @xmath204 the lowest energy soliton is given by @xmath203 . * _ transmutation : _ isospinning hopf solitons can deform into minimal - energy solutions of a type that also exists at @xmath183 ( e.g. @xmath205 ) . * _ formation of new solution types : _ new solution types can emerge which are unstable for vanishing @xmath173 . for example , for @xmath133 the @xmath179 deforms into @xmath182 with the later only being stable for @xmath206 naturally one expects these effects to be present and increasingly relevant for higher hopf charges ( @xmath207 ) since the number of ( local ) energy minima grows with the hopf charge @xmath23 @xcite . in this article we have focussed on purely classically isospinning soliton solutions in the skyrme - faddeev model . the relevance of classically ( iso)spinning soliton solutions was discussed in ref . @xcite in the context of the skyrme model . there it was argued that classically spinning skyrmions could be used to model classically the quantized skyrmion states . for example , a spin-1/2 proton in its spin up state can be interpreted within this approximate classical description as a hedgehog skyrmion of topological charge @xmath208 spinning anticlockwise relative to the positive @xmath45 axis and with its normalized pion fields @xmath209 orientated in such a way that @xmath210 for @xmath211 , respectively . analogously , classically spinning hopf soliton solutions can classically model the quantized spectra of glueballs @xcite . to do this , it is necessary to determine the ( iso)space orientations that describe the excited states of glueballs . to approximate states of non - vanishing spin , rotations in physical space have to be implemented in our computations , which significantly complicates our numerics . our numerical results are of relevance for the quantization of the classical soliton solutions . there are two main methods used in the literature to obtain quantized hopf solitons : the bosonic , semiclassical collective coordinate quantization @xcite and the fermionic quantization @xcite that is based on the finkelstein - rubinstein ( fr ) approach @xcite . both approaches assume that the symmetries of the classical hopf configurations are not broken by centrifugal effects . in the semiclassical bosonic collective coordinate quantization procedure glueballs can be modeled by quantum mechanical states on the moduli space the finite - dimensional space of static minimal energy hopf solutions in a given topological sector which is generated from a single hopf configuration by rotations and isorotations . the effective hamiltonian on this restricted configuration space is canonically quantized . the numerical calculations presented in this paper could be seen as a classical approximation to the collective coordinate dynamics on the moduli space . the allowed quantum states have to satisfy the fr constraints @xcite which follow from the continuous and discrete symmetries of the classical hopf configurations : for a bosonic quantum theory the fr constraints result in constraints for the wave functions defined on the configuration space , whereas fermionic quantization @xcite constrains the wave functions on the covering space of configuration space . ground states and first excited states of hopf solitons for charges up to @xmath212 have been calculated in ref . @xcite using the symmetries of the classical hopf solutions given in ref . it would be instructive to work out the spectra that emerge from the classical solutions calculated in our article . the isospinning , minimal energy hopf solitons of charge @xmath213 are particularly interesting since their symmetries are different from those of the static configurations which are commonly used to calculate the solitons possible ground states . the presentation of a self - consistent , non - rigid quantization procedure goes far beyond the scope of this paper and is the subject of future research . similar results were also reported in a very recent paper @xcite which appeared when our paper was in preparation . the authors in @xcite carried out most of their calculations with @xmath214 and the potential choice @xmath11 . differences to our results are that they neither identify a @xmath190 nor a @xmath198 configuration . unfortunately they did not visualize the linking structure of their 6- and 8-hopf soliton solutions , so that we could not compare them . differences to our results could be due to the different potential choice or to the different choice of the mass parameter @xmath8 . we would like to acknowledge the use of the national supercomputing centre in cambridge . we thank juha jykk , yakov shnir and paul sutcliffe for useful discussions .	we perform full three - dimensional numerical relaxations of isospinning hopf solitons with hopf charge up to 8 in the skyrme - faddeev model with mass terms included . we explicitly allow the soliton solution to deform and to break the symmetries of the static configuration . it turns out that the model with its rich spectrum of soliton solutions , often of similar energy , allows for transmutations , formation of new solution types and the rearrangement of the spectrum of minimal - energy solitons in a given topological sector when isospin is added . we observe that the shape of isospinning hopf solitons can differ qualitatively from that of the static solution . in particular the solution type of the lowest energy soliton can change . our numerical results are of relevance for the quantization of the classical soliton solutions .
cosmic acceleration is one of the most striking discoveries in modern cosmology @xcite . many works have been done to explain this unexpected phenomenon . there are different mechanisms for the universe acceleration . the most popular idea is assuming a new component with negative pressure known as dark energy . other mechanisms contain various modifications of gravitation theory @xcite and violation of cosmological principle @xcite . there are all kinds of dark energy models including quintessence @xcite , holographic dark energy @xcite , quintom @xcite , phantom @xcite , generalized chaplygin gas model @xcite and so on . among these , the simplest and most successful model is @xmath14cdm model @xcite while it has to face some theoretical problems . for example , the `` coincidence '' problem and the `` fine - tuning '' problem @xcite . from this standard model , the flat universe consists of @xmath15 baryons , @xmath16 dark matter , @xmath17 dark energy @xcite . however , we still know little about dark matter and dark energy @xcite . all we can feel about them are their gravitational effects . observationally , they are degenerate . based on these , a unified dark matter model was proposed which assumes that dark matter and dark energy are from a same dark component @xcite . since the equation of state of the unified dark component is unknown , we assume that it has a constant speed of sound @xmath18 which is equivalent to the affine form @xcite . in order to break the degeneracies , different kinds of observational data are used to constrain cosmological model , among which sne ia , cmb , and bao are usually used to constrain cosmological parameters @xcite . recently , other observations are included . for example , the gamma - ray burst @xcite , cluster gas mass fraction @xcite , lensing @xcite and so on . but these processes need to integrate the hubble parameter to obtain the distance scale . the more information about @xmath1 ca nt be embodied . therefore , it is meaningful to study the @xmath1 data directly . the expression of the hubble parameter can be written in this form @xmath19 which depends on the differential age as a function of redshift . based on jimenez et al . @xcite , simon et al . @xcite used the age of evolving galaxies and got nine @xmath1 data . many works based on these nine data was done @xcite . recently , stern et al . @xcite revised these data at 11 redshifts from the differential ages of red - envelope galaxies . moreover , gaztaaga et al . @xcite took the bao scale as a standard ruler in the radial direction , obtained three more additional data . in this paper , we use the markov chain monte carlo ( mcmc ) method to give constraints on the unified dark matter model from the latest data including the observational hubble parameter data . in section 2 , we give a brief introduction of the udm model . in section 3 , we introduce the observational data set we use . the full parameter space using different combinations of data is exhibited in section 4 . finally , we give a conclusion in section 5 . in this section , we give a brief introduction to unified dark matter model . for details of this model , please refer to @xcite . considering the universe is described by the friedmann - robertson - walker metric and assuming a single udm besides baryons and radiation , the friedmann equation is expressed as @xmath21 where @xmath22 represents the hubble constant today , and @xmath23 where @xmath24 , @xmath25 is the critical density today . the subscripts r , b , k and x represent for radiation , baryons , curvature and udm component respectively . each component satisfies the energy - momentum conservation equation @xmath26 for the radiation and baryons , the equations of state are @xmath27 and @xmath28 , respectively . the conservation equation gives @xmath29 and @xmath30 . for the udm , it is considered as an approximation to an barotropic fluid . since the equation of state @xmath31 is unknown , we assume a constant speed of sound @xmath32 which leads to the 2-parameter affine form @xmath33 combined with the conservation equation , we can regard this component as an effective cosmological constant plus a generalized dark matter which has a constant equation of state @xmath4 . the density evolves as @xmath34 where @xmath35 stands for the generalized dark matter , and @xmath2 stands for the effective cosmological constant component , @xmath36 . the friedmann equation now can be expressed as @xmath37 where @xmath38 , @xmath39 and @xmath40 represent dimensionless radiation , baryon matter and curvature today . @xmath2 and @xmath4 are the parameters we introduced . * we notice that when @xmath41 , the udm model are equivalent to @xmath14cdm . in udm model , we can equivalently think that the generalized dark matter may has a non - zero equation of state . naturally , because the property of the effective cosmological constant component is the same as the one in @xmath14cdm , when the property of generalized dark matter is equivalent to the usual dark matter in @xmath14cdm , udm model totally becomes @xmath14cdm . * * the speed of sound characterizes the perturbations to the dark matter density and pressure which are related to the microphysics . since we know litter about dark sector in our universe , the equation of state is introduced to describe the property of dark sector . however , it can not give us the information about the microphysics and other properties . the speed of sound can give information about the internal degrees of freedom . for example , quintessence has @xmath42 . if we can confirm the sound speed that is below the speed of light , we can get further degrees beyond a canonical , minimally coupled scalar field . if the sound speed is small , the perturbations can be detected on a proper scale . this perturbations will change the gravitational potential , thus affecting the geodesic of photons . as a result , the cosmic microwave background will be changed through the integrated sachs - wolfe ( isw ) effect . in order to show the differences between udm mdoel and @xmath14cdm which corresponds to @xmath41 , we give the cmb temperature angular power spectrum in fig . our calculations are based on camb @xcite . * now , we introduce the data set using to constrain the udm model . the observational hubble parameter as a function of redshift has been used to constrain cosmological parameters @xcite . for the observational data , we choose 11 data from @xcite . these data were obtained through the measurement of differential ages of red - envelope galaxies known as `` differential age method '' . the aging of stars can be regarded as an indicator of the aging of the universe . the spectra of stars can be converted to the information of their ages , as the evolutions of stars are well - known . since we can not observe stars one by one at cosmological scales , we take the spectra of galaxies which contain relatively uniform star population . the surveys including the gemini deep deep survey @xcite , vimos - vlt deep survey and sdss @xcite are related to obtaining data . we consider the red - envelope galaxies which are massive with fairly homogeneous star population . moreover , the data can be obtained from the bao scale as a standard ruler in the radial direction known as `` peak method '' @xcite . though this method , we have three additional data @xcite : @xmath44 and @xmath45 , which are independent . plus the measurement of hubble constant @xcite , we totally have 15 data now . the @xmath46 value of the @xmath1 data can be expressed as @xmath47 ^ 2}{\sigma_{i}^2},\ ] ] where @xmath48 is the @xmath49 uncertainty of the observational @xmath1 data . since the sne ia fist revealed the acceleration of the universe , it has been a well - established method for probing dark energy . recently , the supernova cosmology project ( scp ) released the union2 data set including 557 samples @xcite . it covers a redshift region 0 - 1.4 . these data have been widely used to investigate dark energy models @xcite . the distance modules is given by @xmath50+\mu_{0},\ ] ] where @xmath51 is the luminosity distance . it is defined as @xmath52.\ ] ] the marginalized nuisance parameter @xcite for @xmath46 is expressed as @xmath53 where @xmath54 , @xmath55 , @xmath56 , @xmath48 is the 1@xmath57 uncertainty of the observational data . note that the expression @xmath58 which coincides to eq . ( [ chi2sn ] ) up to a constant @xcite , is used in this paper . the constraint results are the same . for cmb , the shift parameters @xmath59 and @xmath60 are important indicators which are related to the positions of the cmb acoustic peaks . these quantities are affected by the cosmic expansion history from the decoupling epoch to today . the shift parameter @xmath59 is given by @xcite @xmath61,\ ] ] where @xmath62 ( hereafter ) , since we have defined the generalized dark matter . the redshift @xmath63 corresponding to the decoupling epoch of photons is given by @xcite @xmath64,\ ] ] with @xmath65 , @xmath66 . moreover , the acoustic scale is related to the distance ratio . it is give by @xmath67/h_0}{r_s(z_{\ast})},\ ] ] where the comoving sound horizon at photo - decoupling epoch is expressed as @xmath68 from wmap 7 measurement , the best - fit values of the data set is given by @xcite @xmath69 where @xmath70 , and the corresponding inverse covariance matrix is @xmath71 with these , we can calculate the @xmath46 by @xcite @xmath72 the baryon acoustic oscillations we use measure the distance - redshift relation at two redshifts . one is in the clustering of the combined 2dfgrs and sdss main galaxy samples at @xmath73 , the other is in the clustering of the sdss luminous red galaxies at @xmath74 . calculated from these samples , the observed scale of the bao are analyzed to constrain the form of the distance scale . the distance scale of bao is defined as @xcite @xmath75\right)^{1/3}.\ ] ] the peak positions of the bao depend on the ratio of @xmath76 to the sound horizon size at the drag epoch . the redshift @xmath77 corresponding to the drag epoch when baryons were released from photos is given by @xmath78}[(1+b_{1}(\omega_{b}h^2)^{b_2})],\ ] ] where @xmath79^{-1}$ ] and @xmath80 @xcite . in this paper , we choose the measurement of the distance radio ( @xmath81 ) at @xmath73 and @xmath74 @xcite . this quantity is given by @xmath82 where @xmath83 is the comoving sound horizon . from the sdss data release 7 ( dr7 ) galaxy sample , the best - fit values of the data set ( @xmath84 , @xmath85 ) @xcite @xmath86 where the corresponding inverse covariance matrix is @xmath87 the @xmath46 value of this bao observation from sdss dr7 can be calculated as @xmath88 where @xmath89 . and 2-@xmath57 contours of parameters @xmath6 , @xmath4 , @xmath2 in a flat universe , for the data sets @xmath1+cmb+bao.,title="fig:",width=755 ] + and 2-@xmath57 contours of parameters @xmath6 , @xmath4 , @xmath2 in a flat universe , for the data sets sne+cmb+bao.,title="fig:",width=755 ] + and 2-@xmath57 contours of parameters @xmath6 , @xmath4 , @xmath2 in a flat universe , for the data sets @xmath1+sne+cmb+bao.,title="fig:",width=755 ] + and 2-@xmath57 contours of parameters @xmath6 , @xmath4 , @xmath2 , @xmath11 in a non - flat universe , for the data sets @xmath1+cmb+bao.,title="fig:",width=755 ] + and 2-@xmath57 contours of parameters @xmath6 , @xmath4 , @xmath2 , @xmath11 in a non - flat universe , for the data sets sne+cmb+bao.,title="fig:",width=755 ] + and 2-@xmath57 contours of parameters @xmath6 , @xmath4 , @xmath2 , @xmath11 in a non - flat universe , for the data sets @xmath1+sne+cmb+bao.,title="fig:",width=755 ] + . @xmath90.,title="fig:",width=604 ] + the statistics we use to constrain parameters is the usual maximum likelihood method of @xmath92 fitting with the markov chain monte carlo ( mcmc ) method . our code is based on cosmomcmc @xcite . the corresponding convergence of the chains r-1 is set to be less than 0.003 which can guarantee the accuracy well . we combine @xmath1 , sne ia , cmb and bao data to give a global fitting on determining the parameters of udm model . the total @xmath92 is given by @xmath93 fist , we consider a flat universe . in order to embody the @xmath1 data , we choose three combinations of data sets and show the 1-d probability distribution of each parameter ( @xmath6 , @xmath2 , @xmath4 ) and 2-d plots for parameters between each other . fitting results from the joint data of @xmath1+cmb+bao are given in fig 1 , the best - value for @xmath4 is @xmath4=@xmath94 . for comparison , we give the results from sne+cmb+bao in fig 2 . the best - fit value is @xmath4=@xmath95 . we can see the observational hubble parameter data gives a more stringent constraint than sne ia in this model . in fig 3 , the total combinations of them result in @xmath4=@xmath5 . from the result , the parameter @xmath4 which stands for the speed of sound of the unified dark sector seems to be vanishing or slightly larger than 0 . so @xmath14cdm model is strongly supported . we further consider a non - flat universe with curvature . for @xmath1+cmb+bao in fig 4 , the results are @xmath4=@xmath96 , @xmath11=@xmath97 . using sne+cmb+bao in fig 5 , @xmath4=@xmath98 , @xmath11=@xmath99 . this results are consist with the situation in a flat universe . the @xmath1 data also gives a much more stringent constraint on @xmath11 . the total joint data in fig 6 give @xmath4=@xmath100 , @xmath11=@xmath12 , which suggest both the speed of sound and the curvature are very little . our results are much more stringent than previous work @xcite . * our results seem to be consistent with @xmath41 using the data sets we select . does that mean there is no need to introduce a new parameter ? we think although the udm model is contrived , the sound speed of the unified dark matter is related to the microphysics of the dark component while equation of state does not . the sound speed leads to the inhomogeneities of dark matter , thus affecting cmb and matter power spectra . with the development of technique , more information in cmb , matter power spectra and so on may distinguish udm model from @xmath14cdm . moreover , we assume a constant sound speed , the whole effective value is very small , however , if when we consider the evolving speed of sound , the results may be changed . * [ cols="^,^,^,^ " , ] [ tab2 ] in this paper , we constrain the unified dark matter with constant speed of sound from the latest data : the newly revised 15 hubble data , type ia supernovae ( sne ia ) from union2 set , baryonic acoustic oscillation ( bao ) observation from the spectroscopic sloan digital sky survey ( sdss ) data release 7 ( dr7 ) galaxy sample , as well as the cosmic microwave background ( cmb ) observation from the 7-year wilkinson microwave anisotropy probe ( wmap7 ) results . we consider two case . in a flat universe , the constraint results are @xmath2=@xmath3 , @xmath4=@xmath5 , @xmath6=@xmath7 . in a non - flat universe , the constraint results are @xmath2=@xmath8 , @xmath4=@xmath9 , @xmath6=@xmath10 , @xmath11=@xmath12 . this constraint results are much stringent than previous work @xcite , which suggest both the speed of sound of the udm and the curvature are very little and a flat @xmath14cdm model is still a good choice . in order to know the effects of the observational hubble data , we also consider the combinations @xmath1+cmb+bao and sne+bao+cmb . from the results , the @xmath1 data gives a better constraint especially for @xmath4 , @xmath11 and @xmath6 than sne ia in this model . we think the @xmath1 data will paly an important role in cosmology in the future @xcite . * acknowledgments * we are grateful to lixin xu for introducing the mcmc method . we also thank hao wang and jun wang for helpful discussions . this work was supported by the national natural science foundation of china under the distinguished young scholar grant 10825313 and grant 11073005 , the ministry of science and technology national basic science program ( project 973 ) under grant no.2012cb821804 , the fundamental research funds for the central universities and scientific research foundation of beijing normal university . riess , et al . j. 116 ( 1998 ) 1009 , [ arxiv : astro - ph/9805201 ] ; s. perlmutter , et al . , astrophys . j. 517 ( 1999 ) 565 , [ arxiv : astro - ph/9812133 ] ; a.c . pope , et al . , astrophys . j. 607 ( 2004 ) 655 , [ arxiv : astro - ph/0401249 ] ; d.n . spergel , et al . , astrophys . j. suppl . 148 ( 2003 ) 175 , [ arxiv : astro - ph/0302209 ] . g. lemaitre , gen . . grav . 29 ( 1997 ) 641 ; r.c . tollman , proc . ( 1934 ) 169 ; h. bondi , mon . not . ( 1947 ) 410 . t. chiba , n. sugiyama and t. nakamura , mnras , 289 ( 1997 ) l5 [ arxiv : astro - ph/9704199 ] ; m. visser , classical and quantum gravity , 21 ( 2004 ) 2603 [ arxiv : astro - ph/0309109 ] ; k.n . ananda and m. bruni , prd , 74 ( 2006 ) 023523 [ arxiv : astro - ph/0512224 ] . z .- h , zhu and m. sereno , astro . astrophys . 487 ( 2008 ) 831 [ arxiv : astro - ph/0804.2917 ] ; s. cao and z .- h , zhu , astro . accepted , [ arxiv : astro - ph/1105.6182 ] . a. lewis and a. challinor , http://camb.info/ r. jimenez , et al . , apj , 593 ( 2003 ) 622 . l. xu and y. wang , plb , 702 ( 2011 ) 114 [ arxiv : astro - ph/1009.0963 ] ; y. gong , x. zhu and z .- h . zhu , mnras , 415 ( 2011 ) 1943 [ arxiv : astro - ph/1008.5010 ] . s. nesseris and l. perivolarropoulos , prd , 72 ( 2005 ) 123519 [ arxiv : astro - ph/0511040 ] . a. lewis and s. bridle , prd , 66 ( 2002 ) 103511 [ arxiv : astro - ph/0205436 ] , + url : http://cosmologist.info / cosmomc/. z.x . zhai , h.y . wan and t.j . zhang , plb , 689 ( 2010 ) 8 ; c. ma and t.j . zhang , apj , 730 ( 2011 ) 74 .	we constrain a unified dark matter ( udm ) model from the latest observational data . this model assumes that the dark sector is degenerate . dark energy and dark matter are the same component . it can be described by an affine equation of state @xmath0 . our data set contains the newly revised @xmath1 data , type ia supernovae ( sne ia ) from union2 set , baryonic acoustic oscillation ( bao ) observation from the spectroscopic sloan digital sky survey ( sdss ) data release 7 ( dr7 ) galaxy sample , as well as the cosmic microwave background ( cmb ) observation from the 7-year wilkinson microwave anisotropy probe ( wmap7 ) results . by using the markov chain monte carlo ( mcmc ) method , we obtain the results in a flat universe : @xmath2=@xmath3 , @xmath4=@xmath5 , @xmath6=@xmath7 . moreover , when considering a non - flat universe , @xmath2=@xmath8 , @xmath4=@xmath9 , @xmath6=@xmath10 , @xmath11=@xmath12 . these give a more stringent results than before . we also give the results from other combinations of these data for comparison . the observational hubble parameter data can give a more stringent constraint than sne ia . from the constraint results , we can see the parameters @xmath4 and @xmath11 are very close to zero , which means a flat universe is strongly supported and the speed of sound of the dark sector seems to be zero .
we start from the following family of interaction hamiltonians in @xmath77 dimensions ( @xmath78 ) , @xmath79 where @xmath80 is a bosonic creation(annihilation ) operator with the canonical commutation relation @xmath81 = \delta(k - k')$ ] , and @xmath82 is a @xmath48-dependent coupling constant . notice that this could be equivalently written , using the unitary transformation @xmath83 and the redefinition @xmath84 , as a coupling with a bosonic field with the following form @xmath85 in our simulation , we will use the former definition for the bosonic field @xmath86 , which is motivated by the specific transmon @xcite implementation that we shall consider . this is not a restriction and we may use a model with coupling constant @xmath87 by considering a flux qubit instead @xcite . in @xmath77 dimensions , scalar fermionic fields are written as @xmath88 where the operator @xmath89(@xmath90 ) that creates fermionic particles(antiparticles ) satisfies the anticommutation rules @xmath91 and @xmath92 . the scalability of this simplified model will be discussed in the last section . in particular , we discuss on the possibility of including spinors in the treatment of the fermionic fields in the simulation . + * discretization and comoving modes * for the purpose of analyzing an interacting theory that may describe fermion - fermion scattering , pair creation , dressed states , and non - perturbative regimes , we introduce comoving fermionic and antifermionic modes . the @xmath14th input comoving modes are defined in the schrdinger picture as follows @xcite @xmath15 where @xmath16 are the @xmath14th fermionic and antifermionic envelopes centered in the momenta @xmath17 and @xmath18 , respectively . these modes create normalizable propagating wave packets when applied to the vacuum which are suitable for describing physical particles , unlike the standard momentum eigenstates which are delocalized over all space . for our purposes we restrict ourselves to orthonormal envelope functions @xmath16 , such that the comoving modes satisfy , at equal times , the anti - commutation relations @xmath19 . the implementation of hamiltonian ( [ complexsupp ] ) in a superconducting circuit setup is a hard problem because it contains an infinite number of both bosonic and fermionic modes . we will be able to mimic the former by using the continuum of bosonic modes appearing in transmission lines or low - quality resonators . in order to deal with the latter , we consider the field fermionic @xmath20 as composed of a discrete , truncated set of comoving modes . this condition allows us to expand the field @xmath20 in terms of two of these new anticommuting modes as a first order approximation , neglecting the remaining anticommuting modes . thus , the fermionic field reads @xmath21 where the coefficients can be computed by considering the anticommutators @xmath22 and @xmath23 as follows @xmath93 and we have considered @xmath20 in the schrdinger picture . henceforth , we shall omit the superindex @xmath94 since we only consider two creation operators . + * hamiltonian in the schrdinger picture * the hamiltonian associated with the proposed quantum field theory model can be rewritten in the light of the previous assumptions . substituting the expressions for the bosonic and fermionic fields of equations ( [ bos ] ) and ( [ coferm ] ) , respectively , into the interaction hamiltonian of equation ( [ complexsupp ] ) yields @xmath95 in order to connect eventually with the circuit simulator , it will prove convenient to use now the jordan - wigner transformation @xcite , that relates the four fermionic operators with tensor products of pauli matrices , @xmath96 , @xmath97 , @xmath98 , @xmath99 , and @xmath100 . this mapping requires to consider the fermionic operators in the interaction picture with respect to the free hamiltonian for the fermions , since in this image the operators do not depend explicitly on time . thus , the interaction term of the hamiltonian can be expressed as @xmath101 we see that @xmath102 presents three kinds of interactions , which are analyzed in the main text : single and two - qubit gates coupled to the continuum , and interactions involving only bosonic modes . the superconducting circuit of our proposal consists of an open transmission line coupled to three tunable coupling transmon qubits ( tcqs ) @xcite and three resonators . one of the resonators couples to two transmon qubits , while the other two are used for individual addressing / readout of these transmons . for pedagogical reasons , we describe the interaction of a single tunable coupling transmon qubit with the open transmission line and a single cavity mode . following the lagrangian description in ref . @xcite , the system of a single transmon coupled to the open transmission line and one resonator is represented by @xmath103 here , @xmath104 , @xmath105 , @xmath106 , and @xmath107 are the node fluxes @xcite depicted in fig . [ figs1 ] , associated with the resonator , the open transmission line , the upper island , and the center island , respectively . additionally , @xmath108 is the flux quantum , @xmath109 and @xmath110 are the capacitance and inductance of the resonator , and @xmath111 and @xmath112 the capacitance and inductance per unit length of the open transmission line . finally , @xmath113 and @xmath114 represent the capacitive coupling of the tcq and the resonator and transmission line , respectively . the hamiltonian of the system is given by @xmath115 , where the terms correspond to the transmon , the resonator , the transmission line and the interaction among them , respectively . in what follows , the subindices @xmath116 and @xmath117 refer to the combinations @xmath118 and @xmath107 respectively . thus the transmon is described as @xmath119 with charging energies @xmath120 , \nonumber \\ e_{c_{-}}&=&-\frac{e^2}{2m}\alpha_{+ } , \nonumber\end{aligned}\ ] ] dimensionless gate voltages @xmath121 and interaction energy @xmath122 with @xmath123 the transmission line and resonator hamiltonians may be written in terms of creation and annihilation operators as @xmath124 where @xmath11(@xmath12 ) and @xmath13 are the creation(annihilation ) operator and the frequency associated with the @xmath48th mode of the open transmission line , respectively . in addition , @xmath50 and @xmath125 annihilate and create excitations of frequency @xmath126 in the resonator . we use the same notation of @xmath127 and @xmath12 to describe the quantum field model and the superconducting circuit proposal . the interaction hamiltonian reads @xmath128 where the coefficients are @xmath129 in this hamiltonian , a coupling between the resonator and the open transmission line has been neglected due to its smallness when compared to the rest of the interactions . the coupling energy between the qubit and the resonator is given by @xmath130 , while the coupling with the transmission line depends on the frequency and is proportional to @xmath131 . the hamiltonian ( [ eqhint ] ) can be expressed in terms of pauli matrices if we truncate the hilbert space of the transmon to the two lowest eigenvalues . we are allowed to perform this approximation due to the anharmonicity of the energy distribution , in which the pair of lowest states may be discriminated from the others . hence , the interaction hamiltonian in equation ( [ eqhint ] ) can be written in terms of pauli matrices in the following general form @xmath132 where @xmath133 stands for the identity operator , @xmath134 with @xmath135 correspond to the pauli matrices , and @xmath51 . the fermionic states will be encoded in the two levels of each qubit . this is done via the mapping of fermionic and antifermionic creation and annihilation operators onto nonlocal spin operators acting on the qubits . we recall the jordan - wigner mapping performed previously , @xmath96 , @xmath98 , and we associate the fermionic operators to the following ones acting on the qubit states , @xmath136 where the states @xmath137 and @xmath138 are the levels of a qubit . with this mapping , the vacuum state corresponds to the state @xmath139 , the state with one fermion is @xmath140 , and the state with one antifermion is @xmath141 . fermion self - interaction may be computed by the probability @xmath142 at time @xmath143 , and pair creation and annihilation may be simulated by the transition probabilities between a state with no fermions into a state with a fermion and an antifermion . the state of the qubits can be detected via standard quantum non demolition measurements . additionally , the average boson population and higher order moments may be measured in the open transmission line via the dual - path technique @xcite . in order to access a full - fledged quantum field theory , we need to take into account spinors in the fermionic field , polarizations in the bosonic field , and other couplings in the field theory side , for instance , @xmath144 , @xmath145 , etc . to achieve this , we add more qubits between the resonator and the open transmission line . in this case , an analogous mapping between the fermionic operators and tensor products of pauli matrices is encoded via the @xmath33-mode jordan - wigner transformation @xmath146 where @xmath31 , @xmath32 , with @xmath33 the total number of fermion plus antifermion modes and @xmath147 the identity operator . since fermionic couplings will appear through bilinears , this encoding will encompass all usual cases . the consideration of bosonic fields beyond one single scalar field may be implemented by the use of multiple open transmission lines .	we propose an analog - digital quantum simulation of fermion - fermion scattering mediated by a continuum of bosonic modes within a circuit quantum electrodynamics scenario . this quantum technology naturally provides strong coupling of superconducting qubits with a continuum of electromagnetic modes in an open transmission line . in this way , we propose qubits to efficiently simulate fermionic modes via digital techniques , while we consider the continuum complexity of an open transmission line to simulate the continuum complexity of bosonic modes in quantum field theories . therefore , we believe that the complexity - simulating - complexity concept should become a leading paradigm in any effort towards scalable quantum simulations . _ _ introduction.__quantum field theories @xcite ( qfts ) are among the deepest and most complex descriptions of nature . this is why different computing approaches have been developed , as feynman diagrams @xcite or lattice gauge theories @xcite . in general , the numerical simulations of qfts are computationally hard , with the processing time growing exponentially with the system size . nevertheless , a quantum simulator @xcite could provide an efficient way to emulate these theories @xcite in polynomial time . for instance , the remarkable developments in superconducting circuits and circuit quantum electrodynamics ( qed ) @xcite , specifically concerning their improvements in controllability and scalability @xcite , make them suitable candidates for developing a quantum simulator @xcite . an important and probably unique property of superconducting devices is that , unlike other quantum platforms , they offer naturally strong and ultrastrong couplings of qubits to a continuum of bosonic modes provided by one - dimensional open transmission lines . for instance , an almost @xmath0 reflection of a single photon by a two - level scatterer in open lines has been demonstrated @xcite , leading to applications of nonclassical state generation of light @xcite . therefore , this system is a specially suited platform to realize quantum simulations of scattering processes involving interacting fermionic and bosonic quantum field theories , where access to the continuum of modes is required . in this letter , we propose the quantum simulation of fermionic field modes interacting via a continuum of bosonic modes with superconducting circuits , by introducing the complexity - simulating - complexity concept . with this we mean using a complex quantum system to simulate another quantum system with similar complexity . along these lines , in our proposal , a continuum complexity in qfts is simulated by a continuum complexity of open transmission lines , instead of approximating the model to a discrete number of modes or reducing it to many qubits . to achieve this goal , we consider a quantum simulator composed of tunable coupling transmon qubits @xcite , an open transmission line with a finite bandwidth of bosonic modes , and a microwave cavity supporting a single mode of the electromagnetic field . in this scenario , we discuss the minimum requirements that superconducting circuits , or any other quantum platform , should fulfill in order to implement a scalable analog - digital quantum simulator , aiming at simulating fermion - fermion scattering , fermion self - interaction , and pair creation and annihilation . in addition , we discuss how to scale up the number of fermionic degrees of freedom for the sake of simulating full - fledged quantum field theories . note that in ref . @xcite , the quantum simulation of a similar quantum field theory model in trapped ions was proposed . however , this quantum platform can only provide a discrete number of bosonic modes that will be hard to improve when considering scalable quantum simulations of qft models . _ the model.__our current understanding of the most basic processes in nature is based on interacting quantum field theories @xcite . for example , models involving interaction of fermions and bosons play a key role . in these kinds of systems , one is able to describe fermion - fermion scattering mediated by bosonic fields , fermionic self - energies , and bosonic polarization . in particular , we will consider a quantum field theory model under the following assumptions : ( i ) 1 + 1 dimensions , ( ii ) scalar fermions and bosons , and described by the hamiltonian ( @xmath1 ) @xmath2 here , @xmath3 is a bosonic field @xcite , with coupling constants @xmath4 , and @xmath5 is the scalar fermionic field , with @xmath6(@xmath7 ) and @xmath8(@xmath9 ) as its corresponding fermionic and antifermionic creation(annihilation ) operators for mode frequency @xmath10 , while @xmath11(@xmath12 ) is the creation(annihilation ) bosonic operator associated with frequency @xmath13 . in order to adapt the simulated model to the simulating setup , we consider a further simplification in eq . ( [ complex ] ) : ( iii ) one fermionic and one antifermionic field comoving modes @xcite interacting via a continuum of bosons . the latter is intended to analyze an interacting theory that may describe fermion - fermion scattering , pair creation , dressed states , and nonperturbative regimes . the @xmath14th input comoving modes are defined in the schrdinger picture as follows @xcite @xmath15 where @xmath16 are the @xmath14th fermionic and antifermionic envelopes centered in the momenta @xmath17 and @xmath18 , respectively . these modes create normalizable propagating wave packets when applied to the vacuum which are suitable for describing physical particles , unlike the standard momentum eigenstates which are delocalized over all space . for our purposes we restrict ourselves to orthonormal envelope functions @xmath16 , such that the comoving modes satisfy , at equal times , the anticommutation relations @xmath19 . the implementation of hamiltonian in eq . ( [ complex ] ) in a superconducting circuit setup is a hard problem because it contains an infinite number of both bosonic and fermionic modes . we will be able to mimic the former by using the continuum of bosonic modes appearing in transmission lines or low - quality resonators . in order to deal with the latter , we consider the field fermionic @xmath20 as composed of a discrete , truncated set of comoving modes . this condition allows us to expand the field @xmath20 in terms of two of these new anticommuting modes as a first order approximation , neglecting the remaining anticommuting modes . thus , the fermionic field reads @xmath21 where the coefficients can be computed by considering the anticommutators @xmath22 and @xmath23 , as follows @xmath24 where we have considered @xmath20 in the schrdinger picture . henceforth , we shall omit the superindices since we only consider two creation operators . the hamiltonian associated with the proposed quantum field theory model can be rewritten in the light of the previous assumptions . substituting the expressions for the bosonic @xmath25 and fermionic @xmath20 fields into the interaction hamiltonian of eq . ( [ complex ] ) yields @xcite @xmath26 the fermionic and antifermionic operators obey anticommutation relations @xmath27 , and the bosonic operators satisfy commutation relations @xmath28\!=\!\delta(k - k')$ ] . in this sense , we expect that reproducing the physics of a discrete number of fermionic field modes coupled to a continuum of bosonic field modes will boost full - fledged quantum simulations of quantum field theories . let us now consider the jordan - wigner transformation @xcite that relates fermionic operators with tensor products of pauli operators : @xmath29 , and @xmath30 , where @xmath31 , @xmath32 , with @xmath33 the total number of fermionic plus antifermionic modes . note that this transformation is efficient with our techniques for simulating fermions coupled to the bosonic continuum . that is , we require a polynomial number of qubits and gates in the number of fermionic modes @xcite . in this case , the hamiltonian in eq . ( [ simple ] ) presents three kinds of interactions : single and two - qubit gates coupled to the continuum @xmath34 , @xmath35 , with @xmath36 for @xmath37 , and interactions involving only bosonic modes , @xmath38 ( see supplemental material @xcite ) . thus , the simulator should provide a mechanism for generating multiqubit gates and coupling spin operators to a continuum of bosons in an analog - digital approach @xcite . in light of the above discussion a possible interaction term reads @xmath39 . the jordan - wigner transformation allows us to write the above interaction as the exponential of a tensor product of pauli matrices with a band of bosonic modes . to compute this exponential , we propose the implementation of the following sequence of quantum gates @xcite @xmath40 } , \!\ ! \nonumber\\\end{aligned}\ ] ] where @xmath41 is a mlmer - srensen gate @xcite that can be parametrized as @xmath42 $ ] . here @xmath43 is extended to as many qubits as fermionic modes are involved , and the central gate @xmath44 is @xmath45 $ ] . + _ _ circuit qed implementation.__circuit qed architectures including the interaction between on - chip coplanar waveguides ( cpws ) and transmon qubits @xcite are an appropriate platform to fulfill the requirements of the analog - digital simulator . we consider the setup depicted in fig . [ fig1](a ) , which consists of a microwave transmission line supporting a continuum of electromagnetic modes ( open line ) interacting with three transmon qubits . in addition , there is a microwave resonator with a single bosonic mode coupled only with two transmons . notice that two superconducting qubits may interact simultaneously with both cpws , while the ancilla qubit interacts only with the open line . in this setup , we consider tunable couplings between each qubit and the cpws , and also tunable superconducting qubit energies via external magnetic fluxes . in particular , the protocol for simulating fermion - fermion scattering will require the ability to switch on / off each cpw - qubit interaction with control parameters . the latter may be realized by combining tunable coupling transmon qubits , @xcite and standard techniques of band - stop filters @xcite applied to the open transmission line . this way , a finite bandwidth of bosonic modes plays a key role in the dynamics . our model considers an open transmission line without an external bath , due to the fact that all dynamical time scales happen before the model is broken by decoherence mechanisms . in this sense , the decoupling of a transmon qubit from the open line may be accomplished by tuning the qubit energy out of the bandwidth . in addition , our protocol may be extended to several fermionic modes by adding more transmon qubits as depicted in fig . [ fig2 ] . in this circuit qed implementation , the system hamiltonian can be written in terms of pauli matrices in the following general form @xcite @xmath46 where @xmath47 is the pauli operator , @xmath11(@xmath12 ) and @xmath13 stand for the creation(annihilation ) operator and the frequency associated with the @xmath48th continuum mode , respectively , whereas the operator @xmath49(@xmath50 ) creates(annihilates ) excitations in the microwave resonator . the coupling strengths @xmath51 and @xmath52 depend on intrinsic properties of the cpw such as its impedance and the photon frequencies . in addition , @xmath53 stands for the @xmath14th qubit position , and the coefficient @xmath54 can be tuned over the range @xmath55([0,\alpha_{\rm{max}}])$ ] via external magnetic fluxes @xmath56 and @xmath57 , which act on the @xmath14th transmon qubit . note that the same magnetic fluxes also allow us to tune the qubit energy . let us discuss how hamiltonian in eq . ( [ hsetup ] ) is able to simulate the dynamics governed by hamiltonian in eq . ( [ simple ] ) . in fig . [ fig1](b ) , we show the set of quantum operations for simulating two - qubit gates coupled to the continuum in a single trotter step @xcite to be realized by the proposed analog - digital simulator . in this circuit qed framework , each gate will correspond to the evolution under the hamiltonian in eq . ( [ hsetup ] ) for specific values of parameters @xmath56 and @xmath57 . specifically , the gates that act on the first two qubits are , from right to left , one mlmer - srensen @xcite interaction @xmath58 , which is mediated by the resonator @xcite , one local gate @xmath59 $ ] that will couple the spin operators to the bosonic continuum , and an inverse mlmer - srensen interaction @xmath60 . the application of these three operations will generate the two - qubit gate coupled with a continuous band of bosonic modes , @xmath61 . the gate @xmath62 will be used independently on each qubit to generate single - qubit gates coupled to the bosonic continuum . besides , the ancilla qubit allows the generation of the gates that involve only the bosonic modes by means of an interaction @xmath63 $ ] , where @xmath64 is the pauli operator . the required gate is obtained by preparing the ancilla in an eigenstate of @xmath65 . the number of entangling gates needed for a single trotter step is eight . the same scheme of gates can be applied on more qubits in order to scale the system for simulating interactions that involve a larger number of fermionic modes . the spatial dependence of the model is given by spatial integrals @xmath66 , where @xmath67 stands for the different space - dependent coefficients appearing in the hamiltonian of eq . ( [ simple ] ) . these integrals can be rewritten as follows @xmath68 \big\ } f(x , t ) . \quad \nonumber \!\!\!\!\!\ ! \\\end{aligned}\ ] ] if @xmath67 satisfies the condition @xmath69 , then we can simplify the integrals such that @xmath70 we can identify the controllable quantity of the circuit @xmath71 with the spatial - dependent terms times the @xmath48-dependent coupling of the field theory model , i.e. , @xmath72 . if we consider an implementation that uses transmon qubits , their capacitive coupling to the open line leads naturally to a coupling @xmath73 , allowing us to simulate models where @xmath74 is constant or weakly dependent on @xmath48 . other kinds of couplings may be simulated by considering a different superconducting circuit such as the flux qubit , leading to the implementation of couplings depending on @xmath75 . _ _ scaling to @xmath33 fermionic modes.__a way of scaling this formalism to a larger number of fermionic modes is to consider more superconducting elements coupled both to the cavity and to the open transmission line , as depicted in fig . [ fig2 ] . if we consider @xmath76 transmon qubits , then , @xmath33 fermionic modes can be also encoded . accordingly , our proposal can implement a large set of fermionic modes interacting with the bosonic continuum . the addition of more qubits and transmission lines will allow one to simulate quantum fields in larger spatial dimensions . this effort would represent a significant advance towards full - fledged quantum simulation of qfts in controllable superconducting circuits . by means of the proposed techniques , one could measure specific features of qfts , such as self - interaction and pair creation and annihilation of fermions mediated via a continuum of bosonic modes . the quantum computation resulting from this quantum simulation is based on unitary evolutions associated with hamiltonian ( [ simple ] ) . this means that at variance with perturbative methods in quantum field theories , the implementation of our protocol will involve an infinite number of perturbative feynman diagrams with a finite number of fermionic modes . in this sense , this approach towards the quantum simulation of full - fledged quantum field theories is significantly different from standard procedures , since it only requires adding more fermionic modes instead of more feynman diagrams . nevertheless , the natural presence of the continuum of bosonic modes in superconducting circuits approaches our proposal to the targeted model . _ conclusions . _ in this letter , we have proposed an analog - digital quantum simulation of fermion - fermion scattering in the context of quantum field theories with superconducting circuits . this quantum technology provides , in a unique and distinct manner , the strong coupling between superconducting qubits and a microwave resonator , as well as between qubits and a continuum of bosonic modes . our approach represents a significant step towards scalable quantum simulations of quantum field theories in perturbative and nonperturbative regimes with a novel approach : simulating the quantum complexity of a continuum of qft bosonic modes with the quantum complexity of a continuum of circuit qed bosonic modes . the authors acknowledge useful discussions with chris wilson , pol forn - daz , michael simoen , and frank deppe . this work was supported by the spanish mineco fis2012 - 36673-c03 - 02 ; ramn y cajal grant ryc-2012 - 11391 ; upv / ehu project no . ehua14/04 ; upv / ehu ufi 11/55 ; upv / ehu phd grant ; basque government it472 - 10 ; ccqed , promisce , and scaleqit european projects and alexander von humboldt foundation . 99 m. e. peskin and d. v. schroeder , _ quantum field theory _ ( westview press , 1995 ) . j. b. kogut , rev . mod . phys . * 51 , * 659 ( 1979 ) ; j. b. kogut and m. a. stephanov , _ the phases of quantum chromodynamics _ ( cambridge university press , cambridge , england , 2003 ) . r. p. feynman , int . j. theor . phys . * 21 , * 467 ( 1982 ) . s. lloyd , science * 273 , * 1073 ( 1996 ) . i. m. georgescu , s. ashhab , and f. nori , rev . mod . phys . * 86 , * 153 ( 2014 ) . t. byrnes and y. yamamoto , phys . rev . a * 73 , * 022328 ( 2006 ) . j. i. cirac , p. maraner , and j. k. pachos , phys . rev . lett . * 105 , * 190403 ( 2010 ) . j. casanova , l. lamata , i. l. egusquiza , r. gerritsma , c. f. roos , j. j. garca - ripoll , and e. solano , phys . rev . lett . * 107 , * 260501 ( 2011 ) . l. mazza , a. bermudez , n. goldman , m. rizzi , m. a. martin - delgado , and m. lewenstein , new j. phys . * 14 , * 015007 ( 2012 ) . j. casanova , a. mezzacapo , l. lamata , and e. solano , phys . rev . lett . * 108 , * 190502 ( 2012 ) . a. mezzacapo , j. casanova , l. lamata , and e. solano , phys . rev . lett . * 109 , * 200501 ( 2012 ) . s. p. jordan , k. s. m. lee , and j. preskill , science * 336 , * 1130 ( 2012 ) . p. hauke , d. marcos , m. dalmonte , and p. zoller , phys . rev . x * 3 , * 041018 ( 2013 ) . v. m. stojanovi , m. vanevi , e. demler , and l. tian , phys . rev . b * 89 , * 144508 ( 2014 ) . a. blais , r. s. huang , a. wallraff , s. m. girvin , and r. j. schoelkopf , phys . rev . a * 69 , * 062320 ( 2004 ) . a. wallraff , d. i. schuster , a. blais , l. frunzio , r.- s. huang , j. majer , s. kumar , s. m. girvin , and r. j. schoelkopf , nature ( london ) * 431 , * 162 ( 2004 ) . i. chiorescu , p. bertet , k. semba , y. nakamura , c. j. p. m. harmans , and j. e. mooij , nature ( london ) * 431 , * 159 ( 2004 ) . h. paik _ et al_. , phys . rev . lett . * 107 , * 240501 ( 2011 ) . m. stern , g. catelani , y. kubo , c. grezes , a. bienfait , d. vion , d. esteve , and p. bertet , phys . rev . lett . * 113 , * 123601 ( 2014 ) . m. mariantoni _ et al_. , science * 334 , * 61 ( 2011 ) . a. a. houck , h. e. treci , and j. koch , nat . phys . * 8 , * 292 ( 2012 ) . d. l. underwood , w. e. shanks , j. koch , and a. a. houck , phys . rev . a * 86 , * 023837 ( 2012 ) . y. chen _ et al_. , nat . commun . * 5 , * 5184 ( 2014 ) . r. barends _ et al_. , nature ( london ) * 508 , * 500 ( 2014 ) . y. chen _ et al_. , phys . rev . lett . * 113 , * 220502 ( 2014 ) . r. barends _ et al_. , arxiv:1501.07703 . o. astafiev , a. m. zagoskin , a. a. abdumalikov jr . , yu . a. pashkin , t. yamamoto , k. inomata , y. nakamura , and j. s. tsai , science * 327 , * 840 ( 2010 ) . i .- c . hoi , c. m. wilson , g. johansson , t. palomaki , b. peropadre , and per delsing , phys . rev . lett . * 107 , * 073601 ( 2011 ) . i .- c . hoi , c. m. wilson , g. johansson , j. lindkvist , b. peropadre , t. palomaki , and per delsing , new j. phys . * 15 , * 025011 ( 2013 ) . i .- c . hoi , t. palomaki , j. lindkvist , g. johansson , p. delsing , and c. m. wilson , phys . rev . lett . * 108 , * 263601 ( 2012 ) . j. m. gambetta , a. a. houck , and a. blais , phys . rev . lett . * 106 , * 030502 ( 2011 ) . s. j. srinivasan , a. j. hoffman , j. m. gambetta , and a. a. houck , phys . rev . lett . * 106 , * 083601 ( 2011 ) . see supplemental material , which includes the full derivation of the system hamiltonian to be simulated , and a detailed derivation for the circuit qed implementation . it also includes refs . [ 8,31,34,40,41,42,43 ] . p. jordan and e. wigner , z. phys . * 47 , * 631 ( 1928 ) . a. altland , and b. simons , _ condensed matter field theory _ ( cambridge university press , cambridge , england , 2010 ) . a. mezzacapo , l. lamata , s. filipp , and e. solano , phys . rev . lett . * 113 * , 050501 ( 2014 ) . k. mlmer and a. srensen , phys . rev . lett . * 82 , * 1835 ( 1999 ) . j. m. chow _ et al_. , nat . commun . * 5 , * 4015 ( 2014 ) . d. m. pozar , _ microwave engineering _ ( john wiley & sons , new york , 2012 ) . j. koch , t. m. yu , j. m. gambetta , a. a. houck , d. i. schuster , j. majer , a. blais , m. h. devoret , s. m. girvin , and r. j. schoelkopf , phys . rev . a * 76 , * 042319 ( 2007 ) . t. p. orlando , j. e. mooij , l. tian , c. h. van der wal , l. s. levitov , s. lloyd , and j. j. mazo , phys . rev . b * 60 , * 15398 ( 1999 ) . m. h. devoret , _ quantum fluctuations : les houches session lxiii _ ( elsevier , new york , 1997 ) . r. di candia , e. p. menzel , l. zhong , f. deppe , a. marx . , r. gross , and e. solano , new j. phys . * 16 , * 015001 ( 2014 ) .
there are many situations where the atmosphere of an object , a star or a planet , is irradiated by a companion star in such a way that this irradiation significantly influences its atmospheric structure . in the case of classical close binary stars , the effect exists , but is rarely dramatic because the effective temperatures differ by a factor of a few . on the other hand , the effect may be quite dramatic in the case of sub - stellar mass objects such as giant planets and brown dwarfs , irradiated by a solar - type star , in which case the ratio of their effective temperatures may reach a factor of 100 or more . we stress that we use the term _ effective temperature _ as used in the theory of stellar atmospheres , namely as a measure of the total energy flux coming from the interior of the object . in the case of extrasolar giant planets ( egp ) , we recently have witnessed a significant increase in interest in predicting egp spectra . this was motivated in part by two detected planetary transits , which are in principle able to provide direct information about the planet s atmosphere . another motivation is the need to predict spectra of extrasolar planets to guide the design of future missions which aim at recording egp spectra . we have recently computed a large set of model atmospheres and spectra of egps ( sudarsky , burrows , & hubeny 2003 - hereafter referred to as sbh ) . we have used a modification of our universal stellar atmosphere program tlusty ( hubeny 1988 ; hubeny & lanz 1995 ) , called cooltlusty . this variant does not compute opacities on the fly ; instead it uses a pretabulated opacity as a function of wavelength , temperature , and density . the tables are computed using an updated version of the chemical equilibrium code of burrows & sharp ( 1999 ) , which includes a prescription to account for the rainout of species in a gravitational field . motivated by the recent discovery of the second transiting planet ogle - tr-56 ( konacki et al . 2003 ; sasselov 2003 ) , which is believed to have a separation of a mere 0.0225 au from its parent star , we tried to extend the sbh models to higher irradiations . however , cooltlusty faced significant convergence problems . after various attempts and using various strategies , we have discovered that the convergence behavior is not the result of a bug in the program , or of an insufficiency in our numerical scheme , but is in fact a consequence of an existence of multiple solutions that may lead to disastrous effects for convergence . when studying the effect , we found that its roots are quite general , and are in fact applicable to other cases in the theory of stellar atmospheres . therefore , we devote the present paper to explain the effect in general terms . the atmospheric structure is obtained by solving simultaneously the radiative transfer equation , the radiative equilibrium equation , and the hydrostatic equilibrium equation . for simplicity , we assume lte . since our basic aim here is to study atmospheres of irradiated giant planets , this approximation is a reasonable one , although it should be relaxed in the future , as was the lte assumption eventually relaxed in modern studies of stellar atmospheres . here , we present a brief overview of the basic equations . the radiative transfer equation is written as @xmath0 where @xmath1 is the specific intensity of radiation as a function of frequency , @xmath2 , angle - described through the cosine of the angle of propagation with respect to the normal to the surface , @xmath3 , and the geometrical coordinate , taken here as the column mass @xmath4 . the later is defined as @xmath5 , where @xmath6 is a geometrical distance ( measured outwards ) , and @xmath7 the mass density . the monochromatic optical depth is defined as @xmath8 . finally , @xmath9 is the source function , given in lte by @xmath10 here , @xmath11 is the true absorption coefficient , @xmath12 the scattering coefficient , and @xmath13 is the total extinction coefficient . all coefficients are per unit mass . the boundary conditions are provided through the diffusion approximation at the deepest point , @xmath14 and the upper boundary condition is @xmath15 where @xmath16 is the specific intensity of the external irradiation . for simplicity , we assume isotropic irradiation , @xmath17 . it is convenient to express the frequency - integrated irradiation intensity as @xmath18 where @xmath19 is an effective temperature of the irradiating star ( in case the irradiation source is not a star , @xmath19 is merely the characteristic temperature of the incoming radiation ) , and @xmath20 a dilution factor . the moments of the specific intensity are defined as @xmath21 \equiv ( 1/2 ) \int_{-1}^1 i_{\nu\mu}\ , [ 1 , \mu , \mu^2]\ , d\mu \ , .\ ] ] the first moment of the transfer equation is written @xmath22 where the second equality follows from equation ( [ sdef ] ) . integrating over frequency we obtain @xmath23 where @xmath24 and @xmath25 are the absorption and planck mean opacities , respectively , defined by @xmath26 and @xmath27 these two opacities are usually assumed to be equal . this is an excellent approximation in the case of non - irradiated atmospheres , because @xmath28 in optically thick regions , and @xmath29 ( the eddington - barbier relation ) , so @xmath30 is proportional to @xmath31 and the averaging over @xmath30 and @xmath31 leads to very similar results . however , here we maintain the distinction because the difference between @xmath24 and @xmath25 turns out to be crucial in the case of strongly irradiated atmospheres . the second moment of the transfer equation is @xmath32 and integrating over frequency we obtain @xmath33 where @xmath34 which is called the flux mean opacity . notice that unlike the two previous opacities , which were averages of the _ true _ absorption coefficient ( without the scattering term ) , the flux - mean opacity contains the _ total _ absorption coefficient . finally , the radiative equilibrium equation is written @xmath35 which can be rewritten , using the above defined mean opacities , as @xmath36 substituting ( [ reint ] ) into ( [ hmomint ] ) , we obtain another form of the radiative equilibrium equation : @xmath37 where @xmath38 is the stefan - boltzmann constant . the above equations are _ exact _ , given lte , but are of course only formal because the mean opacities @xmath24 and @xmath39 are not known in advance ; only @xmath25 is a known function of temperature and density . nevertheless , assuming that they are known and equal to the rosseland mean opacity , we may write a solution for temperature , following the classical textbook procedure , known by the name of `` lte - grey model atmospheres '' ( e.g. mihalas 1978 ) . a generalization of a classical model for the case of external irradiation is given by hummer ( 1982 ) , and for the case of accretion disks and unequal mean opacities by hubeny ( 1990 ) . the procedure is as follows . from equation ( [ reint ] ) we have @xmath40 , which allows us to express @xmath41 through @xmath30 using the well - known relation @xmath42 . to determine @xmath30 , we use the solution for the second moment of the transfer equation @xmath43 , where @xmath44 is the optical depth associated with the flux - mean opacity , and express the moment @xmath45 in terms @xmath30 using the eddington factor , @xmath46 . similarly , we express the surface flux through the second eddington factor , @xmath47 , and we end up with ( see also hubeny 1990 ) @xmath48 + \frac{\kappa_j}{\kappa_b}\ , w\ , t_\ast^4\ , .\ ] ] again , this solution is exact within lte . the usual lte - gray model consists in assuming that all the mean opacities are equal to the rosseland mean opacity . moreover , if one adopts the eddington approximation ( @xmath49 ) , then one obtains a simple expression @xmath50 we will consider the most interesting case , namely that of strong irradiation , defined by @xmath51 . in this case , the second term in brackets is negligible , and we may define a _ penetration depth _ as the optical depth where the usual thermal part ( @xmath52 ) and the irradiation part ( @xmath53 ) are nearly equal , viz . @xmath54 the behavior of the local temperature in the case of the strictly gray model is very simple it is essentially constant , @xmath55 for @xmath56 , and follows the usual distribution @xmath57 in deep layers , @xmath58 . we stress that while the strictly gray model exhibits an essentially isothermal structure down to @xmath59 , such will not be the case for a nongray model , as we shall show in detail in the next sections . in the general case , we have to retain the ratio of the absorption and planck mean opacities . in the irradiation - dominated layers ( @xmath60 ) , the temperature is given by @xmath61 where @xmath62 as stated before , @xmath63 in the case of no or weakly irradiated atmospheres . however , in the case of strong irradiation , @xmath64 may differ significantly from unity , and , moreover , may be a strong function of temperature , and to a lesser extent density . this is easily seen by noticing that in optically thin regions , the local mean intensity is essentially equal to twice the irradiation intensity , because the incoming intensity is equal to the irradiation intensity , and the outgoing intensity is roughly equal to it as well . the reason is that in order to conserve the total flux when it is much smaller than the partial flux in the inward or the outward direction , both fluxes should be almost equal , as are the individual specific and mean intensities . more specifically , @xmath65 if @xmath66 , then we must have @xmath67 for all angles , and thus @xmath68 . we may then write the absorption mean opacity as a function of @xmath41 and @xmath19 as @xmath69 the dilution factor cancels out , and only the spectral distribution of the irradiation intensity matters . in other words , the absorption mean is an average of the opacity weighted by the planck function corresponding to the effective temperature of the source of irradiation , while the planck mean is an average of the opacity weighted by the planck function corresponding to the _ local _ temperature . obviously , they can be quite different . if the monochromatic opacity differs significantly in the region where @xmath70 and @xmath71 have their local maxima , the resulting @xmath24 and @xmath25 will differ substantially . if , moreover , and this is the crucial point , the monochromatic opacity depends sensitively on temperature , the opacity ratio @xmath64 may have a complex , and generally non - monotonic , dependence on temperature . the local temperature in the upper layers is given , in view of equation ( [ tupp ] ) , by the expression @xmath72 it is now clear that if @xmath64 exhibits a strongly non - monotonic behavior in the vicinity of @xmath73 , for instance if it has a pronounced minimum or maximum there , equation ( [ tupp2 ] ) may have _ two or even more solutions _ ! the bifurcation behavior is not limited to the surface layers , but may continue to large optical depths . this might seem surprising at first sight because from equation ( [ tgen ] ) one may expect that once @xmath74 , the mean opacities @xmath24 and @xmath25 become roughly equal , and the eddington factor is @xmath75 , so that @xmath76 all the way till @xmath77 . we recall that the penetration depth may be quite large ; for instance for the case @xmath78 k , @xmath79 k , and @xmath80 ( the case studied in detail in the next section ) , @xmath81 . however , the behavior of the local temperature may be more complicated . this is linked to another interesting inequality of mean opacities that is usually taken for granted , namely the flux mean opacity @xmath39 and the rosseland mean opacity @xmath82 . the rosseland mean opacity is in fact defined in such a way that it is equal to the flux mean opacity in the diffusion approximation . indeed , in this approximation , @xmath83 and therefore @xmath84 where the latter equality represents the definition of the rosseland mean opacity . in the case of strong irradiation , an interesting , and fundamentally different , situation appears . since the net flux is very small , the total flux in the @xmath85 ( outgoing ) hemisphere is roughly equal to that in the @xmath86 ( incoming ) hemisphere ( see above ) . because the monochromatic opacity varies strongly with frequency , there are frequency regions where the net monochromatic flux is positive , and regions where it is negative . the flux - mean opacity close to the surface may attain large values , either positive or negative , depending upon whether @xmath87 weighs the positive or negative net flux regions more . going to deeper layers , the net monochromatic flux decreases because the radiation field becomes more isotropic for all frequencies , so that the flux mean opacity decreases . consequently , the corresponding flux - mean optical depth increments @xmath88 become very small , and @xmath44 will exhibit a plateau where it remains essentially constant with @xmath4 . finally , in the regions where all the influence of the external irradiation dies out , the usual diffusion approximation sets in , and the flux - mean optical depth becomes essentially equal to the rosseland optical depth . this means that for strong irradiation , the temperature in the deep layers should exhibit a plateau with @xmath89 where @xmath90 is the plateau value of the flux - mean optical depth . the value of @xmath90 and thus @xmath91 depends upon the exact form of the monochromatic opacity ; a rough estimate is provided by setting @xmath92 , in which case @xmath93 . using partition functions , lte level densities , stimulated emission corrections , and broadening algorithms , we generated opacity tables from numerous available line lists . for gaseous h@xmath94o , partridge and schwenke ( 1997 ) have calculated the strengths of more than @xmath95 lines . we used a subset of their @xmath96 strongest lines . for various other molecular species ( _ e.g. _ , nh@xmath97 , ph@xmath97 , h@xmath94s and co ) , we used the hitran ( rothman 1992,1998 ) and geisa ( husson 1994 ) databases , augmented with additional lines from theoretical calculations and measurements ( tyuterev 1994 ; goorvitch 1994 ; tipping 1990 ; wattson and rothman 1992 ) . for methane , shortward of @xmath981.0 we used the karkoschka ( 1994 ) opacities and between @xmath981.0 and 1.58 we used the strong ( 1993 ) opacities . longward of 1.58 , we used the dijon methane database ( borysow et al . 2003 ) that includes the hot bands . opacity due to collision induced absorption ( cia ) by h@xmath94 and helium is taken from borysow and frommhold ( 1990 ) , zheng and borysow ( 1995 ) , and borysow , jrgensen , and zheng ( 1997 ) , as updated in borysow ( 2002 ) . feh and crh opacities were taken from dulick et al . ( 2003 ) and burrows et al . ( 2002 ) , respectively . the tio line lists for its nine major electronic systems were taken from the schwenke ab initio calculations , as modified by allard , hauschildt , and schwenke ( 2000 ) . these data include lines due to isotopically substituted molecules @xmath99ti@xmath100o , @xmath101ti@xmath100o , @xmath102ti@xmath100o and @xmath103ti@xmath100o relative to the most abundant isotopic form @xmath104ti@xmath100o . for vo , we used the line list provided by plez ( 1999 ) . because @xmath105v is by far vanadium s most abundant isotope , the lines of isotopically substituted molecules are not necessary . the line lists and strengths for the neutral alkali elements ( li , na , k , rb and cs ) were obtained from the vienna atomic line data base ( piskunov 1995 ) . the general line shape theory of burrows , marley , and sharp ( 2000 ) was used for na and k , while those of nefedov ( 1999 ) and dimitrijevi and peach ( 1990 ) were used for the other alkali metals . in practice , we pre - calculate a large opacity table in @xmath41/@xmath7/frequency space in which we interpolate during the iteration of the atmosphere / spectral model to convergence . the table contains 30,000 frequency points from 0.3 to 300 , uniformly spaced in steps of 1 @xmath106 . generally , 5000 geometrically - spaced frequency points are used in the radiative transfer model . in other words , our approach belongs to the category of `` opacity sampling '' methods for treating line blanketing . we consider two different opacity tables ; one which includes opacity due to tio and vo , and the other one without these molecules . in the case without tio / vo , these species were removed both from the opacity table as well as from the state equation , so these species are considered as completely absent everywhere in the atmosphere . these two different opacity tables suggest themselves because for an irradiated atmosphere we obtain high temperatures in the low pressure outer atmosphere . importantly , temperature inversions are very possible . however , thermochemical equilibrium calculations with rainout ( burrows & sharp 1999 ; lodders & fegley 2002 ) suggest that tio and vo would exist at such high-@xmath41/low-@xmath107 points . since the presence of tio and vo is quite natural at high-@xmath41/high-@xmath107 points , one confronts a situation in which there are two regions rich in tio / vo , separated by a middle region that is tio / vo free . in a gravitational field with a monotonic pressure profile , this gap could act like a cold trap in which the tio / vo that is transported by molecular or eddy diffusion from the upper low - p region into the intermediate cooler region , would condense and settle out , thereby depleting the upper low-@xmath107 tio / vo - rich region . this would eventually leave no tio / vo at altitude to provide the significant absorption that could lead to a bifurcation . however , the cold - trap effect can not be anticipated in abundance tables that are functions of temperature and pressure alone . these tables are not cognizant of gravity , nor are they aware of the global @xmath41/@xmath107 profile . whether this cold - trap effect in fact happens is not yet clear . furthermore , the same cold - trap effect might obtain for other egp species ( e.g. fe ) . hence , we use the presence or absence of tio / vo at high-@xmath41/low-@xmath107 points and the associated ambiguity as a means to explore the real mathematical bifurcation effect we have identified in this paper and leave to a future work the study of the true viability of cold traps in irradiated egp atmospheres . we consider an intrinsically cold giant planet , with @xmath108 k , @xmath109 , irradiated by a g0v star with a spectral energy distribution taken for simplicity to be the corresponding kurucz ( 1994 ) model atmosphere . we consider the class v , `` roaster '' situation , as defined by sbh . the opacity ratio index @xmath64 is plotted in figure [ fig1 ] as a function of temperature , for three densities characteristic to the outer layers of an atmosphere @xmath110 , @xmath111 , and @xmath112 g @xmath113 . the upper panel shows the results for the opacities including tio / vo , while the lower panel shows those without tio / vo . it is clearly seen that in the former case the ratio @xmath114 , and thus @xmath64 , exhibits a sharp peak around 1500 k ; the dependence on density is not strong . the ratios @xmath115 for @xmath116 k ( weak irradiation ) , 1250 k ( medium strong irradiation ) , and 2600 k ( very strong irradiation ) are overplotted with dotted lines . the latter roughly corresponds to the case of the recently discovered transiting planet ogle - tr-56 . according to equation ( [ tupp2 ] ) , the temperature at the surface layers is found at the intersection of the curves of @xmath117 and @xmath115 ; thus , we see that for low and for high irradiation there are unique solutions ( corresponding to @xmath118 k and @xmath119 , but that in the intermediate cases there are indeed multiple solutions . the opacity table without tio / vo does not exhibit a peak around 1500 k , and as it is clearly seen from the lower panel of fig . [ fig1 ] , one obtains unique solutions for the temperature for all reasonable irradiations . mathematically , the case of truly extreme irradiation ( e.g. with @xmath73 around 4700 k ) might again lead to multiple solutions because the line @xmath115 now almost coincides with @xmath117 . however , we do not consider such a case here because then a host of new phenomena which we do not address in our present model , such as departures from lte , dynamical effects , etc . , would likely become crucial . to explain the behavior of the @xmath24 over @xmath25 ratio , we plot in figure [ fig0 ] two examples of the monochromatic opacity , for @xmath120 k ( close to the maximum of @xmath64 ) , and @xmath121 k ( the region near the minimum of @xmath64 ) . in both cases @xmath122 g @xmath113 . we also plot normalized planck functions for the two nominal temperatures , 1600 and 520 k ( dotted lines ) , and for @xmath123 k ( dashed line ) . the latter weighs most heavily the optical region around @xmath124 hz , while the local planck functions put the weight at much lower frequencies ( around @xmath125 and @xmath126 hz , respectively ) . in the case of the higher temperature , @xmath127 k , the monochromatic opacity exhibits a maximum just in the region where the planck function corresponding to @xmath19 is largest ( because of tio and vo ) ; therefore , @xmath128 . for the lower temperature , @xmath129 k , the opacity in the optical range is lower by orders of magnitude , so @xmath24 is now significantly lower . since the planck function for @xmath129 k emphasizes frequencies below @xmath130 hz , where the monochromatic opacity is very large , we obtain @xmath131 . there are two other features seen in figure [ fig0 ] that are worth noticing . first , there is a large increase of @xmath64 for temperatures around and below 100 k. this is a consequence of low opacity at low frequencies . however , the opacity sources are quite uncertain in this regime . for studies of extremely cool objects the opacity table should be checked and possibly upgraded , but studying these objects is not our objective here . second , in the case of no tio / vo opacity ( lower panel ) , it appears that @xmath64 nearly coincides with @xmath132 . taken at the face value , this would imply that the ratio @xmath133 is very close to unity . however , we do not see any compelling reason why this should be generally valid , so we conclude that such a behavior is just a coincidence . in figure [ figtauh1 ] we demonstrate the behavior of the flux - mean opacity discussed in [ temp_deep ] . the high - surface-@xmath41 branch exhibits a negative flux mean optical depth at the surface layers because the opacity is high in the optical region where the next flux is negative . consequently , the flux - mean opacity is negative there , and thus the flux - mean opacity ( and the corresponding optical depth ) on the plateau is lower than in the case of the low - surface-@xmath41 branch . we have also computed exact lte models as described in sbh . we have used the stellar atmosphere code cooltlusty , which is a variant of the universal program tlusty ( hubeny 1988 ; hubeny & lanz 1995 ) . we stress that the emergent flux is not computed by a separate program ; instead it is computed already by cooltlusty . this also means that we do not use different opacity tables for computing the atmospheric structure and for computing the spectra . we have also tested the sensitivity of the computed model on different opacity samplings ; we have resampled our original frequency grid of 5000 points to lower numbers of frequencies , and even for 300 points we found very little differences in the temperature structure ( of the order of a few k ) . in some cases it was found advantageous to converge first a model with 300 frequencies , and using this as an input to reconverge a model with the full grid of 5000 frequencies in the next step . the temperature structure is displayed in figure [ figtemp1 ] , which demonstrates that the simplified description put forward above faithfully reflects the true behavior of temperature . full lines represent models computed for the opacity table without tio / vo , while the dashed and dot - dashed lines represent models with tio / vo . the models without tio / vo always converged to a unique solution , moreover with temperature monotonically decreasing outward , as we may have expected from the above discussion . the behavior of the models with tio / vo is more interesting . the high - irradiation model , the top dashed line , indeed converged to a unique solution . the surface temperature is 2600 k , in excellent agreement with the value expected from figure [ fig1 ] ( the density at the uppermost point is about @xmath134 g @xmath113 ) . even the initial increase of temperature in the upper layers when going inward is explained by figure [ fig1 ] , since for progressively higher densities the solution of @xmath135 occurs at higher temperatures . the low - surface-@xmath41 solution is very close to the no - tio / vo model , which once again demonstrates that the bifurcation is caused by the tio / vo opacity . we have also examined depth - dependent concentrations of tio and vo . indeed , the mole fraction of tio drops precipitously from @xmath136 to essentially zero at @xmath137 g @xmath113 ( that is at pressure @xmath138 bars and temperature @xmath139 k ) for the high - surface-@xmath41 model ; while is is not present in the low - surface-@xmath41 model . behavior of the vo concentration is analogous . the low - irradiation model also converged to a unique solution ; moreover , its surface temperature of about 250 k is in excellent agreement with the value one may expect from figure [ fig1 ] . again , both models , without and with tio / vo agree very well . the two middle curves , the dashed and dot - dashed curves , correspond to the same irradiation , and , thus , represent the bifurcation predicted by our analysis . the model with high surface temperature was converged starting from the extremely irradiated model ( through several intermediate steps ) , while the model with the low surface temperature was converged starting from scratch , i.e. from a usual lte - gray model in which all mean opacities are created equal . the high surface temperature at about 2200 k essentially coincides with the high - temperature intersection of about 2200 k on figure [ fig1 ] . the low - surface-@xmath41 solution exhibits a somewhat higher temperature than would follow from the low-@xmath41 intersection in figure [ fig1 ] 650 versus 550 k , but in view of all the approximations involved such an agreement is still very good . we did not numerically find the third possible solution with a surface @xmath41 of about 1300 - 1400 k. we stress that the numerical solutions were not contrived in any way to lead to the predicted bifurcation . in fact , we first discovered the effect purely numerically we obtained two different , yet perfectly well converged solutions , depending on the input model . we have continued the convergence of both solution until the maximum relative change of temperature and density decreased below @xmath140 in all depth points . several features of the overall accuracy of the model are worth mentioning . in fact , the low value of maximum relative change of temperature and density is not a satisfactory criterion of the convergence . one should also check the accuracy of the computed flux gradient and the value of the net flux . that is , for the former case we have to examine the quantity @xmath141 , and for the latter case the quantity @xmath142 , where @xmath143 is the net flux . for the models presented in figure [ figtemp1 ] , @xmath144 is very small , typically @xmath112 or less . the conservation of the total net flux is much harder to achieve ; the bifurcated models have typically @xmath145 near the surface , while @xmath146 reaches a few @xmath147 for column mass around @xmath148 g @xmath149 ( that is , a few per cent ) , and drops to less than @xmath140 at @xmath150 g @xmath149 and deeper . the error in the net total flux of several per cent might seem large , but one has to bear in mind that the net flux is a subtraction of two large quantities , @xmath151 , and the actual numerical error @xmath152 is of the order of @xmath140 or less . a dramatic demonstration of this fact is that a model , which is almost converged ( maximum relative change in @xmath41 and @xmath7 was less than @xmath153 ) , exhibits a seemingly ridiculous value of @xmath154 at upper layers @xmath155 g @xmath149 ( i.e. an error in the total net flux of some @xmath156% ! ) , yet the actual temperature difference between this and the fully converged model is it most about 2 - 3 k , and the predicted flux in both cases is completely indistinguishable on a plot ! this again shows that in the case of strong irradiation it is the value of @xmath152 ( or @xmath144 ) which is critical for practical convergence , not a much more stringent criterion of the error in the total net flux , @xmath146 . the difference in the temperature structure is of course reflected in the emergent spectrum . in figure [ figflux1 ] we show the emergent flux corresponding to the two solutions of the intermediate - irradiation ( distance 0.08 au ) model , computed with tio / vo . the models differ dramatically in the optical and in near - ir region because of the dramatically different surface temperature and the corresponding increase of the tio / vo opacity . since the low - surface-@xmath41 branch exhibits a higher temperature at the plateau ( see fig . [ figtemp1 ] ) , the flux at the low - opacity regions is higher . finally , fig . [ figflux2 ] displays the emergent flux for the two models for the highest irradiation , namely for the distance 0.0225 au , which corresponds to ogle - tr-56 . the thick line is a model without tio / vo , and the thin line with tio / vo . there is no bifurcation here since both opacity tables led to unique solutions , however quite different ones . the corresponding spectra are thus significantly different as well . we have demonstrated that under certain circumstances the differences between the absorption mean and the planck mean opacities can lead to multiple solutions for the atmospheric structure . this result is quite robust , and in fact represents a so far overlooked general result of elementary stellar atmospheres theory . since we do not expect that the absorption and planck mean opacities will differ significantly in the usual case of radiative equilibrium of non - irradiated atmospheres , the most interesting situations where the effect may play role are strongly irradiated stars and planets , and also , possibly , structures where there is a significant deposition of mechanical energy , such as stellar chromospheres and accretion disks . we are not concerned with these objects here , but would like to stimulate further study of possible bifurcations for these situations . we note that nolan & lunine ( 1988 ) also found a bimodal behavior of the atmosphere of triton which is linked to external irradiation . from the physical point of view we may interpret the absorption mean opacity as the global absorption efficiency , and the planck mean opacity as the global emission efficiency , of the medium . the integrated mean intensity @xmath30 acts as a total absorption pool , while the integrated planck function @xmath31 as a total thermal pool . the radiative equilibrium , that stipulates that the total radiation energy absorbed on the spot is equal to the total energy emitted on the same spot , therefore acts as a thermostat : the radiative equilibrium sets the local temperature in such a way that @xmath157 . here @xmath30 is determined by the radiative transfer , and the local temperature follows as @xmath158 . in the case of strong irradiation , the weighting that determines @xmath24 is dominated by the incoming radiation . if the monochromatic opacity in the region around the peak of the external irradiation is very sensitive to temperature , for instance if it is very high for high temperatures and very low for low temperatures , the `` radiative - equilibrium thermostat '' may find two solutions either a high @xmath41 together with high @xmath24 , or a low @xmath41 with low @xmath24 ; in both cases the radiative equilibrium @xmath157 is satisfied . however , the applicability of this effect to real objects depends critically on a degree of reliability of adopted opacities . for high - temperature conditions where the gaseous opacities dominate , the situation is relatively well under control , thanks to recent enormous progress in computing atomic data ( opacity project , iron project , opal , and others ; for a recent review see , e.g. , nahar 2003 ) . in the case of irradiated giant planets or other substellar - mass objects like l and t dwarfs , the opacity for low density ( @xmath159 g @xmath113 ) and high temperature ( @xmath160 k ) is still relatively uncertain . at the present stage , we can not be sure that the bifurcation effect really occurs , but we have shown that this is certainly a very real possibility which should be taken into account when constructing atmospheric models . the authors are pleased to thank bill hubbard , thierry lanz , jonathan lunine , jim liebert , christopher sharp , and drew milsom for fruitful conversations and help during the course of this work , as well as nasa for its financial support via grants nag5 - 10760 and nag5 - 10629 . we also thank anonymous referee for a number of comments and suggestions . allard , f. , hauschildt , p.h . , and schwenke , d. 2000 , , 540 , 1005 borysow , a. and frommhold , l. 1990 , , 348 , l41 borysow , a. , jrgensen , u.g . , and zheng , c. 1997 , , 324 , 185 borysow , a. 2002 , , 390 , 779 borysow , a. , champion , j.p . , jrgensen , u.g . , wenger , c. 2003 , in stellar atmospheric modeling , eds . i. hubeny , d. mihalas , k. werner , asp conf . 288 , 651 burrows , a. , marley , m.s . , and sharp , c.m . 2000 , , 531 , 438 burrows , a. , ram , r.s . , bernath , p. , sharp , c.m . , and milsom , j.a . 2002 , , 577 , 986 burrows , a. , & sharp , c.m . 1999 , , 512 , 843 dimitrijevi , m.s . and peach , g. 1990 , , 236 , 261 dulick , m. , bauschlicher , c. jr . , burrows , a. , sharp , c.m . , ram , r.s . , and bernath , p. 2003 , accepted to suppl . goorvitch , d. , 1994 , , 95 , 535 hubeny , i. 1988 , comp . 52 , 103 hubeny , i. 1990 , , 351 , 632 hubeny , i. , & lanz , t. 1995 , , 439 , 875 hummer , d.g . 1982 , , 257 , 724 husson , n. , bonnet , b. , scott , n.a . , and chedin , a. 1994 , j. quant . spect . transfer , 48 , 509 karkoschka , e. , 1994 , icarus , 11 , 174 konacki , m. , torres , g. , jha , s. , & sasselov , d.d . 2003 , nature , 421 , 507 kurucz , r. l. 1994 , atomic data for fe and ni , kurucz cd - rom 22 ( cambridge , mass : sao ) lodders , k. , & fegley , b. 2002 , icarus , 155 , 393 mihalas , d. 1978 , stellar atmospheres , 2nd ed . , ( san francisco : freeman ) nahar , s. n. 2003 , in stellar atmospheric modeling , eds . i. hubeny , d. mihalas , k. werner , asp conf . 288 , 651 nefedov , a.p . , sinelshchikov , v.a . , and usachev , a.d . 1999 , physica scripta , 59 , 432 nolan , m.c . , & lunine , j.i . 1988 , geophys . 15 , 1315 partridge , h. and schwenke , d.w . 1997 , j. chem . , 106 , 4618 piskunov , n.e . , kupka , f. , ryabchikova , t.a . , weiss , w.w . , and jeffery , c.s . 1995 , suppl . , 112 , 525 plez , b. , 1999 , private communication . rothman , l.s . , gamache , r.r . , tipping , r.h . , rinsland , c.p . , smith , m.a.h . , benner , d.c . , malathy devi , v. , flaud , j .- , camy - peyret , c. , perrin , a. , goldman , a. , massie , s.t . , brown , l.r . , and toth , r.a . , 1992 , j. quant . transfer , 48 , 469 rothman , l.s . 1998 , j. quant . transfer , 60 , 665 sasselow , d.d . 2003 , apj ( in press ) strong , k. , taylor , f.w . , calcutt , s.b . , remedios , j.j . , and ballard , j. 1993 , j. quant . transfer , 50 , 363 sudarsky , d. , burrows , a. , & hubeny , i. 2003 , , 588 , 1121 tyuterev , vl.g . , babikov , yu.l . , tashkun , s.a . , perevalov , v.i . , nikitin , a. , champion , j .- p . , wegner , ch . , pierre , c. , pierre , g. , hilico , j .- c . , and loete , m. 1994 , j. quant . spect . transfer , 52 , 459 zheng , c. and borysow , a. 1995 , icarus , 113 , 84	we show that under certain circumstances the differences between the absorption mean and planck mean opacities can lead to multiple solutions for an lte atmospheric structure . since the absorption and planck mean opacities are not expected to differ significantly in the usual case of radiative equilibrium , non - irradiated atmospheres , the most interesting situations where the effect may play a role are strongly irradiated stars and planets , and also possibly structures where there is a significant deposition of mechanical energy , such as stellar chromospheres and accretion disks . we have presented an illustrative example of a strongly irradiated giant planet where the bifurcation effect is predicted to occur for a certain range of distances from the star .
recently jain and pati @xcite have shown that dissipation in many - body systems leads to certain geometric phases . it is tempting to examine more closely the conditions associated with the appearance of those phases . it is well known that the low - energy response of many - body systems with spontaneously broken symmetries is dominated by gapless nambu - goldstone ( ng ) bosons . the fact that in jain and pati s approach the geometric phase is related to the static susceptibility and hence to the low - energy response of the system is suggestive of a possible role played by spontaneous symmetry breaking . many - body systems with a broken symmetry of the ground state have certain order parameters and it is natural to expect that the response of a time - dependent external perturbation that couples to an order parameter will cause its change in time . then , following berry s original argument @xcite , one might expect that a cyclic evolution in the ( order ) parameter space can result in a geometric phase . since order parameters in many - body systems are generated by symmetry breaking , no additional parametric dependence of the hamiltonian is needed . it is also well known that the fock space of a many - body system is not unique but splits into folia of inequivalent representations , each labeled by order parameters arising from spontaneous symmetry breaking @xcite . this implies that any temporal evolution of a many - body system involving change of an order parameter takes place within a certain class of inequivalent fock space representations . umezawa and his collaborators working in the formalism of thermo - field dynamics @xcite , @xcite have already shown that temporal evolution through thermally inequivalent vacuua gives rise to dissipation , and vitiello @xcite has specifically pointed out that a geometric phase must be associated with it . all this taken together provides a motivation to examine the conditions that lead to a dissipative geometric phase in a many - body system with spontaneous symmetry breaking . we investigated here these conditions using , as an example , the stoner model of itinerant electron ferromagnetism @xcite . this model has been extensively studied in the past @xcite,@xcite , and various aspects of the mechanism of spontaneous symmetry breaking in it have been elucidated @xcite,@xcite . recently it was applied in the study of isospin excitations in nuclei @xcite . since , to the best of our knowledge , there exist no previous attempts to relate a spontaneously broken symmetry to dissipation and the geometric phase , we consider it worthwhile to study such a relation for a particular case of the stoner model . the results obtained here appear to be generic to systems with broken symmetry . the rest of this paper is organized as follows . in section ii we briefly review the main features of the stoner model at zero temperature following a variational approach which gives the equation for the order parameter ( magnetization ) from minimization of the ground state energy of the system . next , in section iii we show how the emergence of order in the system is related to the appearance of a gapless collective mode , and hence to the low - energy response in the system . both sections are in the nature of an abridged review , and the reader is referred to the original articles of mattuck @xcite and matsumoto et al.@xcite for details . in section iv we study the closed contour integral of magnetization ( the order parameter ) as a function of an the applied external field ( time dependent cyclic perturbation ) and show that it is non - zero if the system is in the symmetry broken phase , and zero otherwise . that essentially gives rise to the dissipative geometric phase of jain and pati in this system and leads to the well known hysteresis curve in magnetic systems @xcite . finally , we extend our results to finite temperature . the hamiltonian of a non - relativistic many - body theory of interacting fermions is generically given by @xmath0 here @xmath1 s are pauli spinors and the two - body interaction potential is su(2)-spin invariant @xmath2 in the stoner model this su(2)-spin invariant electron - electron interaction is treated schematically as a contact - type interaction with a fixed strength . @xmath3 the interaction hamiltonian reduces then to @xmath4 the densities of spin - up and spin - down electrons are related to their respective fermi momenta as @xmath5 while the total electron density @xmath6 remains unchanged , @xmath7 the interaction may change the individual densities of spin - up and spin - down electrons , giving non - zero magnetization @xmath8 @xmath9 it is also useful to define the fermi energy and momentum of unpolarized electron gas by @xmath10 we now evaluate the ground state energy of the system . the kinetic term is given by @xmath11 defining a dimensionless parameter @xmath12 , describing spin polarization ( magnetization ) by @xmath13 one can express the @xmath14 , @xmath15 and @xmath8 as functions of @xmath12 @xmath16 thus one arrives at @xmath17.\ ] ] the potential energy contribution to the ground state energy in hartree - fock appoximation is written as @xmath18 the ground state energy per electron in the hartree - fock approximation ( in units of @xmath19 ) is @xmath20+\frac{\lambda n}{4\epsilon_f}(1-\delta ^2).\ ] ] the value of @xmath12 can be determined by minimizing the ground state energy . @xmath21-\frac{\lambda n}{2\epsilon_{f}}\delta.\ ] ] we arrive at the stoner equation , which determines whether or not the system undergoes spontaneous magnetization @xcite @xmath22.\ ] ] in fig . 1 , we display the plot of magnetization as a function of @xmath23 . for @xmath24 , the system does not develop any spin polarization , i.e. , it remains nonmagnetic . for @xmath25 the system becomes partially magnetized , and for @xmath26 it is fully magnetized . to summarize , in the stoner model when a certain combination of the density and the strength of the interaction reaches a critical value , the system of electrons displays a characteristic ground state instability with respect to spontaneous magnetization.(in this and following figures @xmath8 is used to label the axes , while the actual scale is given in terms of the dimensionless @xmath12 . ) as is well known , the appearance of an order parameter in a quantum many - body system implies the existence of a bosonic degree of freedom in the quasiparticle fock space of this system . this general statement is known as the goldstone theorem @xcite . important as it is , goldstone s theorem turns out to be just one identity out of a chain of the ward - takahashi@xcite,@xcite ( wt ) identities between the exact green s functions of the system . these identities show how the original symmetry of the dynamics of the heisenberg fields is manifested in the properties of the quasiparticles @xcite . here we recapitulate the derivation of these identities according to the paper of matsumoto et al.@xcite . in our non - relativistic field theory of interacting fermions with an @xmath27-spin invariant lagrangian density , we add an external magnetic field , @xmath28 , that couples to magnetization breaking explicitly the @xmath27 symmetry down to @xmath29 . @xmath30 next , we perform a global phase transformation on the fermionic fields @xmath31\psi(x),\ ] ] which we regard as a change of variables in the generating functional of the green s function of our model @xmath32_{h}=\frac{1}{n}\int d \psi d \psi^{\dagger } \exp\ { i\int d^4 x [ l_{h}(x ) + \eta^{\dagger}\psi+\psi^{\dagger}\eta]\},\ ] ] where @xmath33 $ ] . one then notes that , because the functional integral does not depend on the integration variables , the @xmath34-dependence of @xmath35 is fictitious , and hence one should impose the following requirement @xmath36 from this condition follow the basic ward - takahashi identities @xmath37 where @xmath38 is the composite boson field and @xmath39 . by taking appropriate functional derivatives of the basic ward - takahashi identities with respect to the sources , @xmath40 and then setting these sources to zero , one obtains a chain of identities between the exact green s functions in the theory . the lowest order identities in this chain are @xmath41 the first identity shows that the difference in the densities between the spin - up and the spin - down electrons , i.e. the order parameter induced by the external magnetic field @xmath28 , is proportional to the space integral of the propagator of the composite boson field . this boson field is the magnon ( spin - wave ) . it describes the transverse oscillations of the order parameter . the second ward - takahashi identity relates the exact propagators of spin - up and down electrons to their effective coupling to the spin - wave given by the vertex function on the r.h.s of the identity . from our discussion in section ii we already know that for certain ranges of the values of the coupling constant and the density , non - zero magnetization may appear even in the absence of an external field . this means that the l.h.s of the first wt - identity may be non - zero in the limit @xmath42 , which in turn forces the r.h.s to be non - zero in that limit . in the momentum representation this identity is expressed as @xmath43 where @xmath44 is the boson ( magnon ) spectral function , and @xmath45 represents its dispersion relation @xmath46 . this identity implies that in the long wavelength limit , we should have @xmath47 and @xmath48 . in other words a non - zero magnetization in the absence of external field requires the existence of a gapless , ng boson field . this in essence is the goldstone theorem in the context of non - relativistic quantum field theory . this ng boson is physically identified with spin waves or magnons in this system . its propagator has a pole at zero momentum which gives the dominant , low - energy spin response of the system . for a finite magnetic field @xmath49 , the pole contribution to magnetization is given in the momentum representation by @xmath50 with @xmath51 , and @xmath52 . this means that in an external field the magnon becomes massive . this is a common property of pseudo ng bosons@xcite . let us assume now that the external field @xmath28 is changing with time . integrating both sides of equation ( [ pseudong ] ) one gets @xmath53 this essentially agrees with the result of jain and pati @xcite , if one recalls that in the linear response theory the imaginary part of the response function ( the susceptibility ) corresponds to absorption @xcite . in our case all the response at @xmath54 is due to a simple pole in the response function @xmath55 however there is also an important difference . in our appoach it suffices to couple a one - dimensional external field to the system , the second degree of freedom being provided by the order parameter itself . thus no rotation of the external field is needed to generate the geometric phase . this result is satisfactory , since hysteresis in ferromagnets is observed for a magnetic source whose magnitude is varying but whose direction is fixed . now we show that , when the system is in a broken symmetry ground state and the integral ( [ contint ] ) is evaluated for a certain closed contour in @xmath56 plane , that integral is non - vanishing . in order to find the functional relation of magnetization to the external field @xmath57 , we need to evaluate the green s functions of quasiparticles in the system . here we calculate the green s functions approximately by truncating the ( infinite ) chain of wt - identities and solving the resulting system of equations . the lowest order approximation is achieved by truncating the wt - chain of identities at the level of the vertex - propagator identity @xmath58 if one approximates the vertex with a constant @xmath59 , one can determine the approximate boson and fermion propagators consistent with the wt - identities . this corresponds to the hartree - fock ( hf ) approximation for the fermions and the random - phase - approximation ( rpa ) for the boson ( magnon ) field . the approximate fermionic propagators are @xmath60 where @xmath61 is the chemical potenitial of the interacting system , and @xmath62 . note that this structure of the spin - up and down electron propagators follows from the spontaneous @xmath63 symmetry breaking . this structure is observed in experiments as a characteristic splitting in the dispersion relations of the spin - up and down electrons @xcite . having found the green s functions of the single - particle states in hf - approximatation , we now can estimate the densities of spin - up and down electrons . for the majority spin @xmath14 we get @xmath64 and similarly for the minority spin @xmath15 : @xmath65 one needs to determine @xmath66 , and @xmath8 , subject to the following constraints : @xmath67 eqs . ( [ updensity])-([constraints ] ) lead to @xmath68 now using the definitions in eq.([definitions ] ) these conditions can be rewritten as @xmath69 we thus arrive at an equation linking the magnetization to the applied field @xmath70 note that for @xmath71 it reduces to the previously obtained variational condition that determines spontaneous magnetization . thus eq.([orderh ] ) represents a generalization of eq.([stoner ] ) for a case of non zero external field @xmath28 . one can also obtain this condition directly by fixing @xmath28 and minimizing the ground state energy in hf - approximation with respect to @xmath12 @xmath72 in a symmetric ground state , width=240 ] in an asymmetric ground state , width=240 ] the dependence of magnetization on @xmath28 is illustrated in fig.3 and fig.4 for the two cases : a ) the system does not display spontaneous magnetization in its ground state and b ) the system gets spontaneously magnetized in the absence of external field i.e. when @xmath73 . one can relate the characteristic s - shape of the curve @xmath57 to the form of the meandering symmetric valley of the minimum ground state energy . although the bed of this valley remains a local minimum for fixed @xmath28 , this bed does not stay at a constant level but rises and falls as the applied field changes . one can express @xmath28 through @xmath8 from eq.([orderh ] ) , and substitute that expression in @xmath74 . the resulting level of the valley as a function of @xmath8 is given in fig.5 . we see that at the points on the hysteresis curve where @xmath8 jumps ( @xmath75 and @xmath76 ) the destination points have lower energy than the starting points . this indicates that directional tunneling ( in reality also thermally assisted ) between these states may start ahead of the turning points , as often observed in the experimental hysteretic curves . one can see from these pictures that a system that displays an instability of the ground state with respect to spontaneous magnetization , behaves very differently from a system that does not have such instability . the former displays a hysterisis loop when the external field is varied in a cycle , while the latter does not . the area of the hysterisis loop gives the dissipated magnetic energy . in other words we have proven that @xmath77 this area gives the cyclic jain - pati dissipative geometric phase in our model . the result can easily be extended to finite temparature . the dominant part of the low - energy spin response at finite temperature again comes from the magnon pole at @xmath78 , but now its spectral function is temperature dependent @xcite,@xcite,@xcite @xmath79 this suggests that @xmath80 can be determined from a finite temperature extention of eq.([orderh ] ) using the standard technique described in @xcite . we first define @xmath81 then the spin - up electron density is @xmath82 with @xmath83 similarly the spin - down electron density is @xmath84 with @xmath85 defining @xmath86 one finds @xmath87 from which it follows @xmath88 one can in principle determine both @xmath80 and @xmath89 from these equations by solving them numerically . then to compute the finite - temperature berry s phase one simply needs to integrate the finite - temperature order parameter along the hysteresis loop . there is nothing surprising about this result . we have simply come to a reinterpretation of the first law of thermodynamics as a balance of dynamic and geometric phases in a many - body system @xmath90 as often seen before , the geometric phase gives us a new perspective on a classical result . we have shown that when a quantum many - body system with a spontaneous symmetry breaking in the ground state evolves in time in such a way that an order parameter completes a cycle , one can associate a certain geometric phase with the energy dissipated in this cycle . we have examined the stoner model of itinerant ferromagnetism and have shown that when the system has ferromagnetic instability , i.e. , spontaneous magnetization occurs in the absence of applied external field , then hysteresis arises with respect to a changing external field and the magnetization process is irreversible . thus , spontaneous symmetry breaking seems to have bearing upon another important concept in physics - reversibility . the measure of irreversibility is given by the dissipation , i.e. , the dissipative geometric phase . our preliminary investigation has confirmed that a similar mechanism is responsible for the hysteresis observed in other many - body systems displaying spontaneously broken symmetry in the ground state . we plan to address this issue in a forthcoming publication . i am indebted to the late prof . h. umezawa for introducing me to the techniques that he and his collaborators developed and employed to many - body problems over the years . i wish to thank prof . pawel o. mazur for encouraging me to publish this paper . i wish to acknowledge stimulating discussions with the late prof . jeeva anandan , dr . alonso botero , jun suzuki and piruz vargha and to thank them and for their advice . i also wish to thank prof . kuniharu kubodera for a careful reading of the manuscript . this work is supported in part by the national science foundation , grant no . phy-0140214 . 99 s. r. jain and a. k. pati , phys . * 80 * ( 1998 ) 650 m. v. berry , proc . london . a392 ( 1984 ) 457 ; h. umezawa , h. matsumoto and m. tachiki , _ thermo field dynamics and condensed states _ , north - holland , 1982 h. umezawa , _ advanced field theory _ , new - york , 1992 g. vitiello , hep - th/0110182 e. c. stoner , proc . london a * 165 * ( 1938 ) 372 c. herring and c. kittel , phys . rev . * 81 * ( 1951 ) 869 ; c. herring , phys . rev . * 85 * ( 1952 ) 1003 t. moryia , _ spin fluctuations in itinerant electron magnetism _ , springer ser . solid - state sci . * 56 * ( 1985 ) r. d. mattuck and b. johansson , advances in physics 17 ( july 1968 ) 509 ; also in r. mattuck , _ guide to feynman diagrams in the many - body problem _ , dover ( 1976 ) p.299 h. matsumoto , h. umezawa , s. seki and m. tachiki , phys . b * 17 * ( 1978 ) 2276 i. c. danchev , f. c. khanna and h. umezawa , phys . c * 50 * ( 1994 ) 3135 s. lovesey , _ dynamic correlations _ , benjamin / cummings ( 1980 ) g. bertotti , _ hysteresis in magnetism_,academic press ( 1998 ) j. goldstone , nuovo cimento * 19 * ( 1961 ) 154 j. c. ward , phys * 78 * ( 1950 ) 182 y. takahashi , nouvo cimento * 6 * ( 1957 ) 370 ; y. takahashi , _ lecture notes in quantum field theory _ , university of alberta ( 1991 ) r. m. white , _ quantum theory of magnetism _ , springer ( 1985 ) j. p. whitehead , h. matsumoto and h. umezawa , phys . b * 29 * ( 1984 ) 423 a. o. e. animalu _ intermediate quantum theory of crystalline solids _ , prentice - hall inc . , englewood cliffs , new jersey ( 1977 )	we show that the cyclic evolution of an order parameter in a many - body system that has spontaneous symmetry breaking in the ground state brings about a dissipative geometric phase . this phase originates from the same mechanism that leads to dissipation in thermo - field dynamics studied earlier by umezawa and his collaborators and appears to be a manifestation of the dissipative geometric phase introduced recently by jain and pati .
the purpose of the theoretical models is not only to explain current observations , but also to make predictions to be tested with the future observations . in nuclear physics particularly , there still exists a wide gap between experimentally know nuclei and the expected limits of the nuclear landscape , in particularly for the heavier , very neutron rich nuclei @xcite . even though these nuclei are yet to be observed in the laboratory , many of them have impact on astrophysical scenarios @xcite . to predict properties of nuclei , various theoretical models have been developed and applied . in extrapolation to experimentally unknown region , the role of model uncertainties becomes prominent . without accompanied error analysis , the model prediction bears no meaning . presently , the nuclear density functional theory ( dft ) is the only microscopic theory which can be applied throughout the entire nuclear chart . the key ingredient of the dft is the energy density functional ( edf ) , which incorporates complex many - body correlations into a functional , which is constructed from the nucleonic densities and currents . currently , there are three main variants of the nuclear dft ; zero - range skyrme , finite range gogny , and relativistic mean field models @xcite . in this work we focus on the skyrme - edf models . similar kind of analysis for other nuclear dft variants could be also done . similarly like with any other effective theory , the parameters of the nuclear dft models need to be adjusted to empirical input . historically , in the optimization of various nuclear dft models , covariance error analysis of the obtained parameterization was usually neglected . however , as stressed for example in @xcite , error analysis is an essential post - optimization tool to quantify the model errors . only very recently , such kind of analysis has been performed for nuclear dft models @xcite . this kind of information is vital when assessing predictive power of the model . theoretical uncertainties related to the limits of the nuclear landscape , within skyrme - edf models , were studied in @xcite . it was found that , despite large differences in various optimization procedures , current skyrme - edf models give a rather consistent picture about the position of the proton and neutron drip - lines . furthermore , it was found that both , systematic model error and statistical model error , gave rather similar uncertainty width for the drip - line . similar studies for relativistic mean - field models were also done in @xcite . the results were rather similar to skyrme - edf results . uncertainties within the brussels - montreal hfb mass models were analyzed in @xcite . with a given covariance matrix of the model , the propagation of the error from the model parameters to the model predictions can be computed . the information content and effect of a new observable to the uncertainties of the model parameters and predictions was investigated in @xcite . it was found that a precise measurement of neutron skin thickness in @xmath0pb could reduce model errors of other isovector observables . in @xcite propagation of the error within unedf0 edf @xcite model was investigated in semi - magic nuclei . statistical and systematical model errors related to the neutron skin thickness were studied in @xcite . in both of these studies , the poorly constrained isovector parameters were found to be the largest source to the model error . this article is organized as follows : in section [ sec : theory ] we review theoretical formalism . some examples are discussed in section [ sec : results ] , and in section [ sec : conclusions ] we present concluding remarks . in the skyrme - edf framework the total binding energy @xmath1 of the nucleus is @xmath2 where @xmath3 is the kinetic energy density , @xmath4 is the pairing energy density , and @xmath5 is the coulomb energy density . isoscalar ( @xmath6 ) and isovector ( @xmath7 ) time - even part of the skyrme energy density is defined as @xmath8 \rho _ { t}^{2 } + c^{\tau}_{t } \rho_{t}\tau_{t } + c^{\delta\rho}_{t } \rho_{t } \delta \rho_{t } + c^{\nabla j}_{t } \rho_{t } \nabla \cdot { \bi j}_{t } + c^{j}_{t } { \mathbb j}^{2}_{t } \ , , \ ] ] where density dependency is @xmath9 = c^{\rho}_{t0}+c^{\rho}_{t{\rm d}}\rho^{\gamma}_{0 } \ , .\ ] ] in equations above , we have omitted , for brevity , dependence of densities on the spatial coordinate @xmath10 . the skyrme energy density ( [ eq : skyrme ] ) is composed of matter density @xmath11 , kinetic density @xmath12 , spin - current vector density @xmath13 , and spin - current tensor density @xmath14 . isoscalar matter density is defined as @xmath15 , and isovector matter density as @xmath16 , where @xmath17 and @xmath18 are neutron and proton indices , respectively . isoscalar and isovector density is defined similarly also for the kinetic and spin - current densities . standard definitions of these densities can be found e.g. in @xcite . here , @xmath19 \rho _ { t}^{2}$ ] is the local volume term , which gives the largest contribution coming from the energy density to the total binding energy . term @xmath20 is connected to the effective mass . term @xmath21 is the local surface term , which mostly contributes at the surface of the nucleus . furthermore , @xmath22 and @xmath23 are the spin - orbit and tensor terms , respectively . the effect of spin - orbit term ( and tensor term to lesser extend ) is to push spin - orbit partner orbitals apart , to reproduce correct magic numbers . the skyrme energy density of equation ( [ eq : skyrme ] ) is parameterized by the coupling constants @xmath24 . at the present , these parameters can not be pre - calculated from any theory within sufficient accuracy and , therefore , they need to be adjusted to empirical input . there is also a one - to - one correspondence between skyrme coupling constants of the volume part , that is @xmath25 , and infinite nuclear matter ( inm ) parameters @xcite . corresponding inm parameters have been listed on table [ tt : inm ] . inm parameters present a more convenient way to parameterize a part of the edf since their values are approximately known . @ll symbol & explanation + @xmath26 & equilibrium density of the nuclear - matter + @xmath27 & total energy per nucleon at equilibrium + @xmath28 & nuclear - matter incompressibility + @xmath29 & scalar effective mass + @xmath30 & symmetry energy coefficient + @xmath31 & density dependence of the symmetry energy + @xmath32 & vector effective mass + to solve the ground state wave - function , within the single - reference skyrme - edf framework , one has to solve the hartree - fock or hartree - fock - bogoliubov equations self - consistently . this procedure is outlined , e.g. , in @xcite . with the dft models used in nuclear theory , the model parameters are usually adjusted to empirical input , and in some occasions , to some other pseudo - data not directly related to experimental result . typical adjustment involves a minimization of the @xmath33 function , which is usually defined as @xmath34 where @xmath35 is an array of model parameters , @xmath36 is the value of data point calculated from the model , and @xmath37 is the corresponding empirical value for data point . furthermore , @xmath38 is the number of data points , and @xmath39 is the used weight of the data point . the definition @xmath40 is such that it is a unitless function . to obtain the minimum of the @xmath40 , at @xmath41 , various algorithms can be used . in optimization of unedf energy density functionals pounders algorithm was used , which offers significant improvement in terms of used cpu time compared to traditional nelder - mead algorithm @xcite . as for the empirical input , nuclear dft models have been typically optimized to various ground - state properties , see e.g. review in @xcite . in optimization of unedf0 @xcite , binding energies , charge radii , and odd - even staggering data was used . for unedf1 @xcite optimization , data on the fission isomer excitation energies were added to the pool of data points . as a result , fission properties were significantly improved . with unedf2 @xcite , shell structure was addressed by including data on single - particle level energies . the calculated model predictions @xmath36 of equation ( [ eq : chi2 ] ) are , in nuclear dft models , usually non - linear functions with respect of model parameters @xmath35 . in this kind of situation , a rigorous calculation of the covariance matrix is a formidable task , yet to be done . therefore , usually a linearized least square system in the vicinity of the minimum @xmath42 of the object @xmath40 function is assumed . within this approximation , the @xmath36 is assumed to be a linear function of @xmath35 . the validity of this approximation depends on the type of the observable and investigated width of the landscape around the minimum . for example , single - particle energies behave rather linearly with respect of skyrme coupling constants @xcite . also , since the total energy @xmath1 in the skyrme - edf scheme is directly proportional to the coupling constants , as shown in equation ( [ eq : skyrme ] ) , it should also behave relatively linearly . this apparent linearity was also used in @xcite for parameter optimization . nevertheless , non - linearities appear in self - consistent picture , where calculated densities depend on the coupling constants . a linearized approximation to covariance matrix is also numerically more stable compared to alternative choices @xcite . within aforementioned linearized approximation , covariance matrix of the model can be written as a matrix equation @xcite @xmath43_{ij}\ , , \ ] ] where @xmath44 where the weights @xmath39 are the same weights as used in the @xmath40 of equation ( [ eq : chi2 ] ) . derivatives of equation ( [ eq : aij ] ) typically need to be calculated numerically e.g. by taking a finite difference . however , in some cases a closed analytic form may be available . the normalization of @xmath40 of equation ( [ eq : chi2 ] ) and covariance matrix ( [ eq : cov ] ) is selected according to simple convention of @xcite . see the appendix of @xcite for other conventions . closely connected to the covariance matrix is the correlation coefficient between two model parameters , which is defined as @xmath45 where @xmath46 , and @xmath47 is the standard deviation of model parameter @xmath48 . a strong correlation coefficient between two model parameters indicates that the covariance ellipsoid of the @xmath40 landscape , close to the minimum , is strongly elongated @xcite . as an example , figure [ fig : unedf0correl ] shows the correlation matrix of unedf0 parameterization @xcite . from the figure , one can see that some of the volume parameters are strongly correlated with each other , as well as with the isoscalar surface coupling constant @xmath49 . with the covariance matrix known , the standard deviation @xmath50 of an observable @xmath51 can be calculated as @xmath52 where @xmath53 is the predicted value of the observable by the model . again , as with the calculation of the covariance matrix , derivatives with respect to model parameters usually need to be calculated numerically , by taking a finite difference . the standard deviation @xmath50 can be also referred , loosely speaking , as a `` theoretical error '' . however , one should bear in mind , that there are also many other sources of theoretical errors , and it may be very difficult task to estimate the total combined theoretical error . up to so far , we have not defined what the weights @xmath39 of equation ( [ eq : chi2 ] ) should be . the selection of the weights brings some arbitrariness to the adjustment of the model parameters . the position of the minimum @xmath54 depends on the relative magnitude of the used weights between various data points . by changing relative weighting between , e.g. , binding energies and charge radii , the edf model could be tailored to reproduce this kind of data more accurately . as a consequence , in the various edf optimization schemes , there has not been a common practice regarding of selection of the weights . to reduce this kind of arbitrariness , one option would be to use experimental error bars as a weights of the data points in equation ( [ eq : chi2 ] ) . however , in the light of the performance of current edf models , this is hardly a viable solution . experimental mass measurements with penning - traps can reach up to sub - kev level in accuracy @xcite , which is orders of magnitude more precise compared to expected accuracy of roughly @xmath55 of the present day mean - field models . with single particle ( s.p . ) energies , the situation is the same . current skyrme - like functionals can reproduce empirical s.p . data only with a typical root - mean - square ( r.m.s . ) deviation of roughly @xmath55 @xcite . the other alternative is to tune the weights according to the capabilities of the used theoretical framework . as argued in @xcite , the weights should reflect expected accuracy of the model . indeed , this has been the usual choice in many recent skyrme - edf optimization schemes . for example , in optimization of unedf0 , the weights for nuclear binding energy data points were set to @xmath56 , and for the charge radii @xmath57 . this corresponds rather well the overall performance on unedf0 : the r.m.s . deviation for binding energies , over the whole even - even mass table , was @xmath58 and for the charge radii @xmath59 . at the present , due to the lack of any better guideline , it seems that the most appropriate way to set the weights is to use expected accuracy of the model for a each given data type . in this section , some illustrative examples about the propagation of theoretical error are shown . examples are mainly calculated with unedf0 skyrme - edf . however , similar studies can be also done with other edfs where the full covariance matrix of the model parameters is known . one of the most rudimentary observable , which a universal nuclear edf is expected to reproduce well across the nuclear chart , is the binding energy . as discussed in sec . [ sec : theoryweights ] , presently nuclear dft models have a typical r.m.s . deviation from experimental data of the order of @xmath55 . since the calculated total energy @xmath1 , within the skyrme - edf model , is directly connected to the model parameters , the uncertainty of these parameters propagates to the predicted energy . of binding energies together with the absolute values of residuals to the experimental data in er isotopic chain , with unedf0 edf . all in units of mev . ] figure [ fig : beer ] shows the calculated theoretical error in predicted binding energies of even - even erbium isotopes within unedf0 edf model . shown are also absolute values of residuals compared to experimental data of @xcite , that is , the difference between experimental and theoretical result . the results , as well as the results of sec . [ sec : resultsnskin ] , were calculated with computer code hfbtho @xcite . since the volume part of unedf0 is parameterized with inm parameters , as well as corresponding part of the covariance matrix , they are also used here in calculation of derivatives of equation ( [ eq : stddev ] ) . tensor coupling constants @xmath60 in the unedf0 edf model are set to zero . propagation of error to the binding energies of semi - magic isotopic and isotonic chains was also studies in @xcite . similarly as in @xcite , the propagated error here increases when moving further out from the valley of stability . relation of the model residuals to experimental data was also investigated for lead isotopic chain , and it was found that around double - magic @xmath61 , the residuals were much larger compared to propagated error . this , and the fact that residuals along the isotopic chains are nor randomly distributed , but usually show arc - like features ( see e.g. @xcite ) , are clear examples of the deficiency of the model . here , in er isotopic chain , the difference between residuals and propagated error is not as drastic , due to the fact that this kind of skyrme edf models are expected perform better in deformed open shell nuclei . , but for the two - neutron separation energies . ] theoretical error for two - neutrons separation energies is shown in figure [ fig : s2ner ] along with the residuals to experimental values for unedf0 . as can be seen , the magnitude of residuals and the magnitude of theoretical error is usually rather similar . this is an indication that the the weights selected for unedf0 optimization are in balance with expected accuracy of the model . the only notable deviation exists at the magic neutron number @xmath62 . theoretical error for two - neutron separation energy was also studied in @xcite for sv - min @xcite edf . similarly as with sv - min , the error increases towards neutron rich regime . the same was also found in @xcite . recently , there has been a considerable theoretical and experimental interest on the neutron skin thickness . neutron skin is defined as a neutron matter distribution extending further out compared to proton matter distribution . neutron skin can be characterized by its thickness , which correlates strongly with other isovector observables @xcite . it has also a strong connection to neutron matter equation of the state @xcite . experimentally , various probes have been used to determine neutron skin thickness . a recent measurement , the lead radius experiment ( prex ) @xcite , determined the parity - violating asymmetry coefficient in @xmath0pb , which is directly related to the neutron skin thickness providing @xmath63 @xcite . this is probably the most model independent measurement of the neutron skin thickness . error propagation to predicted neutron skin thickness was studied in @xcite , where both , systematic and statistical , model error were addressed . one of the major outcomes of @xcite was that the statistical model error is the defining theoretical uncertainty in the case of neutron skin thickness , whereas systematic model error was found to be notably lower . figure [ fig : nskin - ca - sn ] shows predicted neutron skin thickness in even - even calcium and tin isotopes together with theoretical statistical error and error budget . both , neutron skin thickness and related theoretical uncertainty , increase towards more neutron neutron rich isotopes . as shown on the lower panels , the density dependence of the symmetry energy , @xmath31 , has the largest contribution to the total error of the neutron skin thickness in both , calcium and tin isotopes . the situation was found to be the same also for lead isotopes in @xcite . the second largest contributor comes from symmetry energy , @xmath30 , similarly again as was found in @xcite . this is one example , among many others , about the poorly determined isovector parameters in the edf models . . label [ pie12 ] refers to systematic model average of @xcite . unedf0 , sv - min and systematic model error are taken from @xcite . all in units of fm . ] current status related to the uncertainties of the neutron skin thickness in @xmath61 is summarized in figure [ fig : nskin208pb ] , which shows the result of prex experiment together with systematic model error of @xcite , and the systematic and statistical model error obtained in @xcite . the statistical model error is shown for unedf0 and sv - min edfs . as discussed in @xcite the prex measurement , with a rather large error bar , can not constrain current edf models . planned prex - ii @xcite and crex @xcite experiments , which will measure neutron skin thickness in @xmath61 and @xmath64ca , aim at much higher experimental accuracy . results provided by these experiments could help to constrain isovector parameters in the future edf optimization schemes . in this article , we have analyzed theoretical uncertainties related to the nuclear edf models . in particularly , the focus has been on the propagation of the error to the predicted observables within the framework of unedf0 skyrme edf . nuclear binding energies , two - neutron separation energies and neutron skin thickness were used as an examples . in all cases , as well as in the previous studies , it was found that theoretical error increases towards increasing isospin excess . this has been directly linked to the poorly known isovector parameters of the edfs . similar kind of analysis could be done for other edf models , where information about the covariance matrix exists . nevertheless , with this kind of analysis , one should always bear in mind , that the procedure outlined in this article can only assess theoretical uncertainties within the model itself . uncertainties and model errors connected to the deficiency of the model itself can not be estimated with this procedure . one signature of such kind of deficiency is a clear pattern of residuals of experimental data , which are significantly larger compared to calculated theoretical model error . in @xcite binding energies close to double - magic @xmath61 , and single - particle energies were found to follow this kind of pattern . yet another source of errors are the used numerical and computational procedures . in the future edf parameter optimization schemes , sensitivity analysis of the obtained parameterization will be prime importance . without such a procedure , there is no handle to judge the predictive power of the model . as concluded in @xcite , the limits of the skyrme - like edfs have been reached and major improvements are not to be expected on this path . to go beyond present day edf models , many alternatives have been presented @xcite . nonetheless , all of them share the common feature of the requirement to adjust the model parameters to empirical input . it will be exciting to see how these new edf models perform across the nuclear chart and how good will be their predictive power . this work was supported by the academy of finland under the centre of excellence programme 2012 - 2017 ( nuclear and accelerator based physics programme at jyfl ) and under the fidipro programme , and by the european union s seventh framework programme ensar ( thexo ) under grant no . we acknowledge the csc - it center for science ltd , finland , for the allocation of computational resources . 10 m. kortelainen , j. mcdonnell , w. nazarewicz , e. olsen , p .- g . reinhard , j. sarich , n. schunck , s. m. wild , d. davesne , j. erler , and a. pastore . nuclear energy density optimization : shell structure . , 89:054314 , 2014 . s. e. agbemava , a. v. afanasjev , d. ray , and p. ring . global performance of covariant energy density functionals : ground state observables of even - even nuclei and the estimate of theoretical uncertainties . , 89:054320 , 2014 . m. stoitsov , m. kortelainen , s. k. bogner , t. duguet , r. j. furnstahl , b. gebremariam , and n. schunck . microscopically based energy density functionals for nuclei using the density matrix expansion : implementation and pre - optimization . , 82:054307 , 2010 . stoitsov , n. schunck , m. kortelainen , n. michel , h. nam , e. olsen , j. sarich , and s. wild . axially deformed solution of the skyrme - hartree fock bogoliubov equations using the transformed harmonic oscillator basis ( ii ) hfbtho v2.00d : a new version of the program . , 184:1592 , 2013 . m. warda , x. vias , x. roca - maza , and m. centelles . neutron skin thickness in the droplet model with surface width dependence : indications of softness of the nuclear symmetry energy . , 80:024316 , 2009 .	parameters of the nuclear density functional theory ( dft ) models are usually adjusted to experimental data . as a result they carry certain theoretical error , which , as a consequence , carries out to the predicted quantities . in this work we address the propagation of theoretical error , within the nuclear dft models , from the model parameters to the predicted observables . in particularly , the focus is set on the skyrme energy density functional models .
precision calculations for the lhc and future colliders are entering more and more the domain of predictions beyond the next - to - leading order ( nlo ) . the rapid progress was mainly triggered by the development of better calculational methods and tools , but also by the development of more powerful computing resources . both aspects lay the groundwork allowing to move towards the automation of calculations beyond nlo . in these proceedings we will focus on the automation of two - loop calculations , in particular two - loop integrals with several mass scales . while the latter are very difficult to calculate analytically , numerical methods may offer a solution . however , the numerical integration can only succeed if the various types of singularities which may occur in a two - loop amplitude can be isolated or evaded beforehand , preferably in an automated way . the program secdec @xcite is able to perform this task , and to subsequently integrate the resulting finite parameter integrals numerically . in the following we will describe the program secdec and some of its new features , as well as its embedding into a framework to generate two - loop amplitudes . , width=377 ] the steps to be performed in order to generate a virtual amplitude ( at in principle any loop order ) in an approach based on feynman diagrams are roughly the following : 1 . generation of the contributing diagrams in terms of algebraic expressions , based on the feynman rules derived from a lagrangian for a certain model , 2 . casting the expression into a form `` coefficients @xmath0 ( tensor ) loop integrals '' , 3 . reduction of the loop integrals to a set of master integrals ( ideally , this set is a minimal basis in loop integral space ) times coefficients , where the latter coefficients will depend on the dimension @xmath1 of loop momentum space , 4 . evaluation of the master integrals , combination with the coefficients forming the amplitude . we are constructing a program package which is designed to perform the steps listed above in an automated way . it partly builds on the program package gosam @xcite for automated one - loop calculations ; for example it also uses qgraf @xcite for the diagram generation . however , the reduction step and the evaluation of the master integrals are much more challenging than at one loop , as the integral basis is usually not known beforehand , and the master integrals may not be available in analytic form . within the gosam-2loop framework , we constructed an interface to the reduction program reduze @xcite . the reduction to master integrals is then run separately , and the resulting database is imported back into gosam-2loop . the latter also contains an interface to the program secdec which is used to evaluate the integrals numerically . the flowchart for gosam-2loop , shown in fig . [ fig : gosam2 ] , illustrates the main steps . the program secdec@xcite is based on the sector decomposition algorithm described in @xcite . it can be applied to multi - loop integrals as well as more general parametric integrals which contain dimensionally regulated end - point singularities . other public implementations of sector decomposition are described in @xcite . we consider scalar multi - loop integrals in order to simplify the notation . the case of tensor integrals is analogous , for details we refer to @xcite . a scalar feynman integral @xmath2 in @xmath1 dimensions at @xmath3 loops with @xmath4 propagators , where the propagators can have arbitrary , not necessarily integer powers @xmath5 , has the following representation in momentum space : @xmath6 where the @xmath7 are linear combinations of external momenta @xmath8 and loop momenta @xmath9 . introducing feynman parameters and carrying out the momentum integrations leads to @xmath10 with @xmath11 . the graph polynomial @xmath12 is a positive - semidefinite function of the feynman parameters only , while @xmath13 is a function of feynman parameters and the kinematic invariants characterising the integral . in the region where all lorentz invariants formed from external momenta are negative , which we call the _ euclidean region _ , @xmath13 is also a positive semi - definite function of the feynman parameters @xmath14 and the invariants . if some of the invariants ( masses , squared external momenta ) are zero , the vanishing of @xmath13 may lead to an infrared ( soft or collinear ) divergence . thus it depends on the _ kinematics _ and not only on the topology ( as in the ultraviolet case ) whether a zero of @xmath13 leads to a divergence . if some of the kinematic invariants are positive , @xmath13 can also vanish at points in feynman parameter space which are not the endpoints of the parametric integration . to be able to deal with such thresholds ( or pseudo - thresholds ) , a deformation of the integration contour into the complex plane has been implemented @xcite . the program secdec can construct the graph polynomials @xmath12 and @xmath13 automatically . acting on polynomials raised to some power which usually contains the dimensional regulator @xmath15 , it will extract the poles in @xmath16 and evaluate their coefficients numerically , as explained in the next subsection . the program consists of two main parts , one being designed for loop integrals , to be found in a directory called loop , the other one for more general parametric integrals , located in a directory called general . the procedure to isolate the poles in the regulator @xmath15 and to do the subtractions and integrations is very similar in the two branches . however , only the loop part contains the possibility of contour deformation , because only for loop integrals the analytic continuation can be performed in an automated way , following feynman s @xmath17 " prescription . the basic flowchart of the program is shown in fig . [ fig : structure ] . . [ fig : structure ] , scaledwidth=70.0% ] secdec-3.0 @xcite has a number of new features , aiming more towards applications in which the program is used as part of the calculation of a full virtual amplitude , rather than serving only as a numerical check for master integrals which are calculated analytically . among these features are : * the usage on a cluster has been facilitated . this has been achieved by decoupling the parts where mathematica is needed completely from the numerical part , and by creating a directory structure such that the files for the numerical integration created by secdec can be transferred to a cluster , where they are compiled and run . run scripts are provided for the pbs , lsf and condor batch systems . ( the scripts for the lsf system are only available in the most recent version 3.0.9 . ) + the kinematics can be defined in a simple text file kinem.input containing the numerical values for the invariants occurring in the integral , where the program will loop over all kinematic points defined in kinem.input . this new way to define the kinematic points to be calculated is also available in single machine mode . * integrals with numerators can be calculated not only as contracted tensor integrals , but also by defining the numerators as inverse propagators . further , the list of propagator powers can also contain zero entries . these features allow an easy interface to integral reduction programs as e.g. reduze @xcite , where the propagator list defines the integral family , while the list of propagator powers defines the considered integral within this family . * two new decomposition strategies , based on algebraic geometry @xcite , have been implemented . these strategies by construction can not run into an infinite recursion at the iterated decomposition . remarkably , the strategy g2 often leads to a lower number of subsectors than the `` default '' strategy x. this new feature is described in detail in @xcite . * the masses contained in the propagators can also be complex . for details we also refer to @xcite . * integrals with propagators which are only linear in the loop momentum can be calculated as well . if these integrals are non - euclidean , care has to be taken to ensure the correct sign for the contour deformation . the program will assume the same @xmath18 prescription as for feynman propagators which are quadratic in the loop momenta . * a new treatment of spurious poles which have a worse than logarithmic divergence behaviour ( related e.g. to factors of @xmath19 in the decomposed integrand ) is available , which uses integration by parts ( ibp ) to remap these poles to logarithmic poles . this option is invoked if the parameter ibpflag is nonzero . * the user can define thresholds below which the imaginary part of the integral is known to be zero ; in this case secdec will not attempt to calculate the imaginary part . this can speed up the calculation considerably . * for the numerical integration , the user can also choose to use an integrator from mathematica . for one - dimensional integrals , the program automatically switches to cquad @xcite , being faster and more accurate than monte carlo integration . * the user interface has been facilitated , decoupling completely the following three types of input : ( a ) definition of the graph name , desired expansion order in @xmath15 and parameters for the numerical integration ( param.input ) , ( b ) graph definition ( math.m ) , ( c ) definition of the kinematics ( kinem.input ) . the next release of secdec will also contain the possibility to create a library containing the sector functions produced by the program , such that the integrals can be linked to the calculation of a full amplitude . this will allow to adjust the precision to which an individual integral is calculated according to its importance within the full amplitude . finally , for secdec-4.0 , the algebraic part will be completely restructured , and will be implemented in ` python ` rather than mathematica . this will be described in more detail in the following subsection . the forthcoming implementation of the algebraic part of secdec in ` python ` has several advantages . one of them is certainly given by the fact that ` python ` is an open source software . in addition , the new implementation offers much more modularity . the various steps the program performs , like construction of the graph polynomials , iterated decomposition , subtraction , @xmath15-expansion ( see also fig . [ fig : structure ] ) , are available in the form of ` python ` modules and can be called separately . the decomposition can act on products of polynomials raised to some power , so it is not limited to a loop integral structure of the form @xmath20 times a numerator ( see eq . ( [ eq : fu ] ) ) . further , the new implementation allows to perform an expansion in several regulators , not only the dimensional regulator @xmath15 . the additional regulator could stem for example from analytic regularisation @xcite or from parameter integrals containing the string tension @xcite . an alpha - version of the python implementation can be found at http://secdec.hepforge.org , where a manual is also linked . as an example for an integral where the analytic result is unknown , we give results for a non - planar two - loop four - point function ( see fig . [ fig : hhnp2b ] ) containing both internal masses ( @xmath21 ) and massive external legs with a different mass ( @xmath22 ) . the result for this diagram is of the form @xmath23 . numerical results in the form of an interpolated scan over 20 values for @xmath24 times 20 values for @xmath25 for the real part of the finite term @xmath26 are shown in fig . [ fig : resultshhnp2b ] . ) . the dashed lines denote massive legs with a mass ( @xmath22 ) different from the internal mass . [ fig : hhnp2b ] , scaledwidth=40.0% ] , for the values @xmath27173gev , @xmath28125gev , @xmath29 . [ fig : resultshhnp2b ] , scaledwidth=75.0% ] we have presented new features of the program secdec , which is a program to factorise poles from dimensionally regulated parameter integrals , and to evaluate the pole coefficients numerically . the embedding of secdec into gosam-2loop , a general framework for the calculation of 2-loop amplitudes , has also been briefly described . the new features of secdec and further developments of these programs aim at the usage of secdec within the calculation of full two - loop virtual amplitudes in cases where the analytic result for most of the master integrals is unknown , rather than the calculation of just a few kinematic points to check an analytic calculation of individual integrals . this development is in line with the rapid developments in high performance computing . 10 url # 1#1urlprefix[2][]#2 carter j and heinrich g 2011 _ comput.phys.commun . _ * 182 * 15661581 ( _ preprint _ ) schlenk j and zirke t 2016 calculation of multi - loop integrals with secdec-3.0 _ proceedings , 12th international symposium on radiative corrections ( radcor 2015 ) and loopfest xiv ( radiative corrections for the lhc and future colliders ) _ ( _ preprint _ )	we report on the development of tools to calculate loop integrals and amplitudes beyond one loop . in particular , we review new features of the program secdec which can be used for the numerical evaluation of parametric integrals like multi - loop integrals .
over the past several decades , a large number of investigations of the stars and the globular clusters in the galactic halo have concluded that the halo formed chaotically and over a long period of time through the accretion and disruption of dwarf galaxies ( e.g. , searle & zinn 1978 , norris 1986 ; carney et al . 1990 ; majewski 1992 ; mackey & gilmore 2004 ; see also freeman & bland - hawthorn 2002 for a review ) . there is evidence , however , that the inner and outer parts of the halo have sufficiently different properties that at least two mechanisms are required to explain the halo ( e.g. * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? accretion appears to have dominated the formation of the outer halo , but the flattening and prograde rotation of the inner halo and the old ages of the globular clusters in the inner halo suggest that its formation involved the dissipative merging of gas clouds at an early epoch ( see * ? ? ? * and references therein ) . while this dual halo picture is supported by substantial evidence , it continues to be important to see what fraction of the halo can be explained by , for example , an accretion - only scenario . the most direct evidence for accretion comes from the detection of halo substructures consisting of co - moving streams of stars . the best documented of these are the streams originating in the sagittarius ( sgr ) dwarf spheroidal ( dsph ) galaxy , which cross much of the sky ( * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * among others ) . several other substructures have been recently detected in the halo of the milky way ( * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * among others ) and in the halo of m31 @xcite . these streams provide strong evidence in favor of the hierarchical model of the formation of galaxies , in which large galaxies such as the milky way formed through multiple mergers with smaller building blocks . simulations of the hierarchical picture predict that the milky way has experienced a large number of minor mergers ( see for example * ? ? ? * ) and show that the ones that occurred during the past few gyrs . may be recognizable today as stellar streams . while these simulations appear to be consistent with the current observations of the outer halo @xcite , the search for more streams and better documentation of the known ones remains important . we report here an investigation of the halo substructure in the direction of virgo . the first sign of an overdensity of halo stars in this direction came from the quest rr lyrae star ( rrls ) survey @xcite , which revealed a small overdensity near 13 hr right ascension ( @xmath4 ) , which was the west boundary of the survey as it existed then . as the survey was extended , it became clear that this feature is considerably larger and is centered near @xmath5 hr @xcite . in the jargon of the quest survey it became known as the _ `` 12.4 hr clump '' _ ( declination ( @xmath6 ) was not specified because of the narrow range of the survey around @xmath7 ) . the rrls in the clump cover a range of distances from the sun ( @xmath8 ) with the greatest concentration at @xmath9 kpc @xcite . using photometry from the sdss , newberg et al . ( 2002 ) independently discovered an overdensity in f type main - sequence stars in approximately the same direction and at a similar distance , which they called s297 + 63 - 20.5 . they suggested that it may be related to the small group of rrls identified earlier by @xcite , and several authors have subsequently suggested that the 12.4 hr clump of rrls and s297 + 63 - 20.5 are the same substructure or at least related . also using the sdss photometric catalogue , juric et al ( 2005 , 2008 ) later identified a very large overdensity of main sequence stars toward virgo ( the _ virgo overdensity _ ( vod ) ) . on the sky , the vod encompasses both the 12.4 hr clump and s297 + 63 - 20.5 . but , according to @xcite , the vod is spread along the line of sight from @xmath10 to @xmath11 kpc , which is a larger range than either the 12.4h clump or s297 + 63 - 20.5 . the spectroscopy of rrls in the 12.4 hr clump by @xcite provided the first evidence of a moving group in the virgo direction , which they called the _ virgo stellar stream _ ( vss ) . the recent spectroscopy of f type main - sequence stars in s297 + 63 - 20.5 by @xcite revealed a rich structure in velocity space , but with a major peak that nearly coincides with the velocity of the vss . they tentatively conclude that s297 + 63 - 20.5 , the vss , and the vod are likely associated , if not the same . on the other hand , several authors @xcite have argued that the virgo substructures are parts of sgr streams , while others @xcite have rejected this hypothesis and suggested instead that they are due to a separate merger event involving a dwarf galaxy . we are carrying out a program to observe spectroscopically a large number of quest rrls in order to measure their radial velocities and metallicities . rrls are arguably the halo tracer that can be most precisely located in space , and with the addition of radial velocities of even modest precision ( @xmath12 @xmath13 ) , they become powerful probes of phase space . this wonderful property of rrls comes with a price , for they are expensive observationally . complete light curves are necessary to obtain both mean magnitudes ( hence , precise distances ) and ephemerides . the later are necessary to isolate the motion of the star in the sky from the pulsational velocity of the star by means of the fitting of a radial velocity curve . in this paper , we present spectroscopic observations of bright rrls ( @xmath14 , equivalent to d @xmath15 kpc ) in the part of the quest survey that overlaps with the vod . after examining these data for halo substructure , we discuss the relationships between our results and the previous detections of substructures in virgo and also consider the possibility that they are related to streams from the sgr dsph galaxy . the rrls surveys by @xcite and by @xcite provide important data on parts of the vod that are not covered by the quest survey . according to @xcite , the vod occupies @xmath16 sq . degrees of the sky , and its rough limits are shown in figure [ fig - zone ] . notice that the sdss does not extend south of @xmath17 at @xmath18 ( see figure 6 in * ? ? ? thus , it is possible that the vod extends farther in that direction . the region surveyed by quest is a @xmath19-wide band centered at @xmath20 . it passes over part of the vod , as shown in the figure [ fig - zone ] . the dot - dashed line in figure [ fig - zone ] shows the southern limit of a second quest catalog ( vivas et al . in preparation ) which is centered at @xmath21 . eight stars from this second region are included in our sample . the region in common between the vod and the quest survey is roughly limited by galactic longitude in the range @xmath22 . at the declination of the quest survey , those limits are equivalent to @xmath23 . we have therefore selected for spectroscopy all rrls in the quest catalog @xcite within these limits in @xmath4 , and with magnitude @xmath24 . the constraint in magnitude was imposed by the limitations of our major instrument , the smarts 1.5 m telescope ( see [ sec - observations ] ) . the magnitudes of this sample correspond to a range of distance from the sun of @xmath25 to @xmath26 kpc ( using @xmath27 , following vivas & zinn 2006 ) , which partly overlaps the range estimated for the vod ( @xmath28 kpc , * ? ? ? figure [ fig - quest ] shows the extinction corrected magnitude @xmath29 ( equivalent to distance ) of part of the quest rrls in the northern galactic hemisphere , as a function of @xmath4 . the box encloses the sample of 44 stars studied in this work . within the box , all the stars in the first quest catalog were observed spectroscopically , but only 8 rrls ( out of 38 ) from the second quest catalog were observed . for reference , we have indicated with solid symbols the location of the rrls that probably belong to the vss @xcite . the stars in our sample are located in the same area of the sky as the vss , but are closer to the sun . most of them ( the ones with @xmath30 kpc ) are located within the same region of the sky and distance range as the vod . notice also in figure [ fig - quest ] that part of the tidal stream of the sgr dsph galaxy lies in the background , at @xmath31 kpc . we obtained a total of 88 spectra of the 44 rrls with 3 different telescopes between 2001 and 2007 . the instrumental setups of the different observing runs are summarized in table [ tab - telescopes ] , and the individual observations are detailed in table [ tab - observations ] . most of the spectra were obtained with the smarts 1.5 m telescope at cerro tololo interamerican observatory , chile ( denoted as `` smarts '' in table [ tab - observations ] ) . we also used observations obtained with the 1.5 m telescope in la silla , chile ( denoted as `` eso '' ) and with the hydra multifiber spectrograph at wiyn in kitt peak national observatory , usa . most of the spectra were taken at blue wavelengths , covering the balmer lines ( beginning in h@xmath32 ) and the ca ii h and k lines . however , we took advantage of available time from other projects at the wiyn telescope and obtained spectra using a setup which covered the red part of the spectra , including the ca triplet lines . the wiyn observations are denoted in table [ tab - observations ] as `` wiyn - b '' and `` wiyn - r '' for blue and red spectra respectively . for each target star , we tried to obtain at least two spectra taken at different phases during the pulsation cycle . thirty five out of the 44 rrls have two or more spectra available . table [ tab - observations ] contains i d ( same quest i d as in vivas et al . 2004 ) , julian date , exposure time , telescope , phase of observation ( see [ sec - ephemerides ] ) , heliocentric radial velocity with its error ( see [ sec - rv ] ) , and metallicity ( see [ sec - feh ] ) . the 8 stars from the unpublished second part of the quest catalog have i d numbers higher than 700 . the spectrum of star # 254 indicates that is too cool to be a rr_c _ star . this is not surprising because a small fraction of the quest rr_c _ stars are expected to be misidentifications of other types of variable stars , mainly variable blue stragglers and w uma eclipsing binaries ( see * ? ? ? star # 254 was excluded from any further analysis . to avoid excessive broadening in the spectral features due to the changing velocity of the star during the pulsational cycle , exposure times were kept short , usually under 30 minutes . a few stars have exposure times as long as 60 minutes , which is equivalent to @xmath33 of the pulsation cycle of a typical rrls . in all observing runs , we observed several stars from the list of @xcite which are both radial velocity and pseudo - equivalent width standards . since good ephemerides are essential for fitting the radial velocity curves , it is important to improve them as much as possible . the quest survey has a large number of epochs ( an average 32 for our targets ) , which should yield precise ephemerides . in @xcite , the time of maximum light ( hjd@xmath34 ) for the rr_c _ stars was set by the observation with the brightest magnitude , which is not the best procedure because it can be skewed by photometric error and by poor sampling near maximum light . in addition , for some rr_c _ stars , it was necessary to list two possible periods because both produced reasonable light curve shapes to the eye . to obtain better light curve parameters , we have fitted a light curve template to each rr_c _ star in our sample . a sine curve produced a poor fit because a typical rr_c _ stars has a phase difference between maximum and minimum light of @xmath35 . consequently , we constructed a light curve template based on 9 of the best observed quest rr_c _ stars , which did not suffer from period aliases or from a poorly measured hjd@xmath34 . this template was then fitted using @xmath36 minimization , while varying the amplitude , phase at maximum light , magnitude of maximum light , and period around the values listed in the quest catalogue . we also explored the parameters around the @xmath37 alias periods . the ephemerides for the rr_ab _ stars in the quest catalogue were more secure because they were already determined from light curve fitting and because period aliases were less of a problem for them . however , the period was not allowed to vary in the fitting . this time it was allowed to vary in the fitting , but the resulting light curves are not significantly different from the old ones . only in 5 cases were the changes in period equal to or slightly larger than @xmath38 days . the ephemerides that yielded the best fits to the light curves for the rr_c _ and rr_ab _ stars have been adopted here and are listed in table [ tab - velocities ] . the metallicities of the rrls have been measured from the spectrograms following the methodology and calibration of @xcite , which is based on freeman & rodgers ( 1975 ) modification of the @xcite @xmath39s method . the method , which involves plotting the pseudo - equivalent width of the ca ii k line ( w(k ) ) , corrected for interstellar absorption , against the mean pseudo - equivalent widths of the @xmath32 , @xmath40 , and @xmath6 balmer lines ( w(h ) ) , is described in detail in @xcite . in general , the type c rrls are less numerous than the type ab in old stellar populations , e.g. globular clusters , and because they are harder to detect than type ab , their number in rrls surveys of the field suffer more from incompleteness . @xcite did not produce a [ fe / h ] calibration for type c rrls , which were ignored in his study of the metallicities and kinematics of field rrls . there is a sufficient number of type c stars in our sample that it is important to measure [ fe / h ] from our spectrograms . the average type c variable is @xmath41k hotter and has a @xmath42 dex larger surface gravity than the average type ab variable , although the types overlap in both quantities , particularly when the type ab variables are hottest during their pulsations . the layden calibration may therefore require no modification when measuring type c rrls . this is supported by kemper s ( 1982 ) demonstration that the @xmath39s method , which was devised for type ab variables , can be applied to type c , and more directly by @xcite , who found that both types of variables in same globular cluster define a tight sequence in a plot of k line strength against hydrogen line strength with some intermingling of the types along the sequence . however , @xcite defined their k line and h line strengths differently than did @xcite . to see if layden s calibration can be applied to the type c , we observed both type c and type ab variables in the globular cluster m3 with the hydra multiobject spectrograph on the wiyn telescope . while the fiber setup targeted a large number of variables of both types in this variable - rich cluster , the combination of partly cloudy weather and variable fiber throughputs limited the number of usable spectra to 8 type c stars and 10 type ab . analyses of high dispersion spectrograms of red giants @xcite and rrls @xcite and the tightness of the red giant branch in the color - magnitude diagram of m3 indicate that there is very little if any variation in [ fe / h ] among the stars in this cluster . for our purposes , we assume that there is no real star to star scatter and investigate whether layden s calibration yields the same [ fe / h ] when applied separately to the two types of rrls . our measurements for the m3 rrls in the layden pseudo - equivalent width system are plotted in figure [ fig - feh ] , where different symbols are used to depict the type ab and type c stars . the mean [ fe / h ] obtained from all 10 of the type ab variables is @xmath43 with a standard deviation of 0.27 . the large scatter is undoubtedly due to the low s / n of some of the spectrograms . the 5 type ab rrls ( filled squares in figure [ fig - feh ] ) for which we have the highest s / n spectrograms ( @xmath44 at the caii k line ) yield @xmath45[fe / h]@xmath46 with a more reasonable std . dev . of 0.14 . the 8 type c variables yield @xmath45[fe / h]@xmath47 ( std . @xmath48 ) , and the 5 that have spectrograms with @xmath49 yield @xmath45[fe / h]@xmath50 ( std . dev . @xmath51 ) . the high s / n samples differ in @xmath45[fe / h]@xmath52 by @xmath53 , which suggests that the offset , if any , between the type ab and c variables is small . since these mean values agree well with zinn & west s ( 1984 ) result for m3 ( [ fe / h ] = -1.66 with an internal precision of @xmath54 ) , layden s calibration appears to be equally applicable to type c and type ab rrls . the metallicities that we have derived from the spectrograms of our program stars are listed in the last column of table [ tab - observations ] . the majority of the spectrograms have @xmath55 , and from the experiences with the m3 variables , we estimate that the precision obtained from one spectrogram is @xmath56 for the type ab variables and @xmath57 for the type c. the lower precision for the type c rrls is consequence of the convergence of the constant [ fe / h ] lines at large w(h ) ( see figure [ fig - feh ] ) . a few spectrograms have @xmath58 , and the [ fe / h ] values derived from them are marked with colons in table [ tab - observations ] . layden s calibration breaks down for type ab variables that are observed on the rising branch , where the strength of the k line is reduced by the relatively large effective gravity and shock waves may produce emission in the cores of the h lines that reduce their equivalent widths . we have discarded measurements [ fe / h ] that were made at phases where these effects are likely to be important . the surface gravity variations of the type c variables are milder , and the equivalent widths of the h lines are not affected by emission . none of their measurements have been discarded . radial velocities for all blue spectra were obtained by fourier cross - correlation ( iraf s _ fxcor _ task ) with radial velocity standard stars of a and f spectral type , in the spectral range @xmath59 . the radial velocity standards were separated in two groups according to their effective temperature . each target star was cross - correlated with 6 to 14 different standard spectra of the group with more similar effective temperature . for each spectrum , we obtained a mean radial velocity , @xmath60 , by weighting the result from each standard by the cross - correlation error returned by _ fxcor_. this error depends on both the s / n of the spectra and the similarity between the target star and the standard . in order to estimate the radial velocity errors produced by the different telescope - spectrograph combinations , we cross - correlated each spectrum of the radial velocity standards with the spectra of the other standards , taking care to avoid cross - correlating two spectra of the same star . reassuringly , the differences between the mean velocities from these correlations and the literature values were close to zero for all of the standards with each instrumental setup . we adopted the standard deviations of these differences as the minimum error in velocity that we can achieve with a particular setup for our target stars ( 16 , 8 , and 12 km / s for the smarts , wiyn blue , and eso 1.5 m spectra , respectively ) . because these errors include the uncertainties in the zero - points of the wavelength calibrations , we believe that these values better reflect the true errors than the means of the cross - correlation errors from _ fxcor _ , which are smaller . cross - correlation was not used to calculate the radial velocities of the red hydra spectra because the subtraction of the strongest night - sky lines left residuals in some wavelength regions . instead , gaussian line profiles were fitted to the unblended paschen lines of hydrogen and/or the two strongest lines of ca ii , which however at the resolution of the spectrograms are each blended with a paschen line . the ca ii lines were only measured in spectra that were taken at the coolest phases of the type ab variables when the paschen lines are much weaker than the ca ii lines . before computing the radial velocities from the ca ii lines , we made slight adjustments to their wavelengths to take into account the blending . these were estimated from high s / n spectrograms of bright stars of known velocity whose spectra resemble the type ab variables at minimum light . to test this method , we obtained with the smarts 1.5 m telescope red spectra of 6 bright type ab rrls during cloudy weather . the velocities obtained from the measurements of the ca ii lines did not systematically deviated from the ones measured from the paschen lines alone , and the systemic velocities ( see below ) obtained for these stars agreed with values in the literature . for the red hydra spectra , the standard deviation of the mean of the velocities given by the individual lines ranged from 4 to 10 @xmath13 , which is consistent with the results obtained from similar spectra of bright stars of known velocity . the best technique for obtaining the radial velocity of the center of mass ( the systemic velocity , @xmath61 ) of a type ab rrls is to integrate the radial velocity curve given by weak metal lines over the whole pulsational cycle , and the results obtained for rrls in globular clusters ( e.g. * ? ? ? * ) are consistent with the radial velocities of the clusters to within the internal velocity dispersions of a few @xmath13 . in our low resolution spectrograms , the weak lines produce poorly determined peaks in the cross - correlations , whereas well - defined ones are produced by the balmer lines and other strong features . furthermore , we could not obtain sufficiently large numbers of spectra to determine the velocity curve for each star . for the rr_ab _ , we adopted instead the method described by @xcite , which involves measuring the velocities at a few phases ( @xmath62 ) from low resolution spectrograms and then fitting them with the radial velocity curve of the ab variable x arietis that @xcite obtained from measurements of @xmath63 . observations made near maximum light ( @xmath64 and @xmath65 ) were not used in these fits because the radial velocity curve undergoes a large discontinuity in this region ( see for example * ? ? ? * ) . the velocity curve that is determined from the h lines has a larger amplitude and differs in other ways from the one given by the metals lines ( see * ? ? ? * ; * ? ? ? nonetheless , layden s ( 1994 ) comparisons of his measurements of @xmath61 with literature values for the same stars indicates that this method produces results that are consistent with the errors of the measurements and the fits of the velocity curve . in the case of the template star x ari , the systemic velocity is equal to the h line velocity at @xmath66 , but there is some small star to star variation in this value . in our calculation of the errors in @xmath61 , we have included terms to account for this variation and for the variation in the amplitudes of the radial velocity curve ( see * ? ? ? * ) . for rr_c _ stars , which have much smaller velocity amplitudes than the rr_ab _ , there are at most small offsets between the velocity curves given by the metal lines and the h lines . to measure @xmath61 , we fitted a template constructed by combining the radial velocity curves of t sex and dh peg ( see * ? ? ? * ) . figure [ fig - lc ] shows the best fit for each star having two or more observations . these plots include all of the radial velocity measurements for a star , even if they were not used in the fits because they are close to @xmath67 . in each case , these points lie within the expected range , which provides a check that our measurements do not suffer from large systematic errors . notice also that six of the stars were observed with two different telescopes , and the consistency of the measurements , as seen in figure [ fig - lc ] , is also an indication that large systematic errors are not present . in the fitting of the radial velocity curves , the data were weighted by their errors . the mean difference between the data points and the radial velocity curve , @xmath68 , in most cases was of the same size , or smaller , than the typical errors in the individual measurements ( see table [ tab - velocities ] ) . for the rr_c _ , we calculated the error in the systemic velocity by combining the individual @xmath69 in the usual manner for calculating the @xmath70 of a weighted mean value . for both types of stars , if @xmath68 was larger than @xmath71 we adopted the former as our final error in the systemic velocity . the mean error in the systemic velocities of all , but one , rrls range from 7 to 28 @xmath13 , with a mean value of 17 @xmath13 . one star , the type c # 174 , had an unusually bad fit of the radial velocity curve and its final error is 42 @xmath13 . finally , we calculated the radial velocity measured by an observer at the sun who is at rest with the galactic center , @xmath72 by @xmath73 our results are presented in table [ tab - velocities ] , which lists : i d , coordinates @xmath4 and @xmath6 , the extinction corrected mean v magnitude , the type of rr lyrae star , the period and julian date at maximum light ( see [ sec - ephemerides ] ) , distance from the sun ( d ) , number of spectra available and number of spectra actually used in the fitting of the radial velocity curve , the error in the fit of the radial velocity curve ( @xmath68 ) , the systemic velocity ( @xmath61 ) and its error ( @xmath71 ) , and the velocity in the galactic rest frame , ( @xmath72 ) . as noted above , the vod encompasses the entire range in @xmath4 and perhaps a larger range of distance than our sample of rrls . the spatial distribution of our sample of rrls ( see the box in figure [ fig - quest ] ) does not reveal any regions of exceptionally high density that are strong candidates for halo substructure . there is a roughly uniform band of stars that crosses the region at distances between 7 and 10 kpc and also a region of relatively high density near @xmath74 kpc within @xmath75 . this second feature is the near side of the large over - density in rrls that we have referred to as the `` 12.4 hr clump '' @xcite , which at 19 kpc contains the vss @xcite . figure 12 in @xcite shows the location of the 12.4 hr clump in the first band of the quest survey , but only two of the stars identified there as probable clump members are in our sample . the addition of rrls from the 2nd quest band makes the clump even more prominent , and there are 9 stars in our sample that now appear to be outliers of the clump . if the band of stars at @xmath76 kpc and the 12.4 hr clump are real halo substructures , many of the stars should have similar values of @xmath72 . figure [ fig - vgsr_histo ] is a histogram of @xmath77 of the whole sample of rrls . the mean velocity is @xmath78 @xmath13 , and the standard deviation is @xmath79 @xmath13 . these values disagree with recent measurements of samples of bhb stars in the halo ( see below ) . for a sample of 1170 bhb stars , @xcite found a line - of - sight velocity dispersion , @xmath80 @xmath81 , and @xmath82 @xmath13 . @xcite found a similar @xmath83 in their sample of bhb selected from the 2mass survey , although they noted that the velocity dispersion increased to @xmath84 @xmath13 if they selected only the most metal poor stars in their sample to avoid thick disk contamination . the shapiro - wilk test for normality @xcite indicates that the @xmath72 distribution of our sample does not deviate significantly from a normal distribution . it seems safe , therefore , to use the student t - test and the f - test , which reject with @xmath85 confidence that our sample has the same mean velocity and velocity dispersion as the bhb sample of @xcite . however , the velocity dispersion of our sample is not significantly different from the value @xcite obtained for their most metal - poor bhb stars . while these differences may be another sign that rrls and bhb stars in the halo have different kinematics ( see below and * * ) , the halo substructure in our sample is definitely a major contributor to its relatively large velocity dispersion and its positive mean velocity . while not statistically different from a normal curve , the histogram in figure [ fig - vgsr_histo ] is clearly unexpected for a random selection of halo objects , for it has a very prominent peak at @xmath86 @xmath13 . in total , there are eight rrls , or 19% of our sample that have @xmath87 @xmath81 , which is very unlikely to be random occurrence ( 1 in 10,000 , if @xmath80 @xmath13 and@xmath82 ) . this excess of rrls with large positive @xmath72 is most likely due to a halo substructure . other likely substructures , which have less extreme values of vgsr , are revealed when distance and velocity are used in concert . in figure [ fig - bin ] , we present histograms of @xmath72 for these 3 distance intervals , where we have selected dividing lines that avoid putting stars on the boundaries . the histograms in figure [ fig - bin ] do not resemble the normal curves expected of random selections of halo stars , and in each case the shapiro - wilk test rejects normality with greater than 98% confidence . when considering the @xmath88 kpc interval , it is important to recall that quest survey is probably incomplete at the bright magnitudes of these relatively nearby stars because of saturation of one or more detectors in the quest mosaic camera ( see * ? ? ? detector saturation was not a problem for the stars in the more distant intervals , and the quest survey is expected to be @xmath89 and @xmath90 complete for the type ab and type c variables , respectively . for these intervals at least , the departures of the @xmath91 distributions from the expected halo curves are therefore probably due the presence of halo substructure . to identify possible groups of stars , we implemented with only a few modifications the algorithm described by @xcite ( the stellar pair search ) . in this technique , the distances of each star from all others in the sample are calculated from the cartesian galactic coordinates @xmath92 and @xmath93 . pairs of stars are ones that are separated by less than a critical distance , @xmath94 and have a difference in radial velociticy less than a critical velocity , @xmath95 . pairs having one star in common form a group . a group may be formed by many connected pairs , and the minimum number of stars in a group , @xmath96 , is 3 ( two pairs connected by one star ) . to study the statistical significance of the groups we performed monte carlo simulations of random samples of stars having the radial distribution of the rrls in the halo . specifically , we used the density radial profile described in @xcite for the direction @xmath97 , @xmath98 , which has a slope of @xmath99 and takes into account the flattening of the inner halo . we assigned to each simulated star a random @xmath4 and @xmath6 within the same limits of our sample , and a random @xmath77 drawn from a gaussian distribution with @xmath80 @xmath13 and @xmath82 . each simulated sample was introduced in the group detection algorithm described above . the experiment was repeated many times , and we counted how many times the simulated samples produced groups with characteristics similar to the groups in the real sample of rrls . the choices for the parameters @xmath94 and @xmath95 require some discussion . as we described above , the vod covers the whole range of right ascension of the sample studied here , which is @xmath100 . any substructure in our data may be as large as this . at 7 kpc from the sun , @xmath101 is equivalent to 5 kpc , and at 12 kpc from the sun , it is 8 kpc . on the other hand , the uncertainty in distance of the rrls is @xmath102 @xcite , which suggests a lower limit of @xmath103 kpc . the streams from either destroyed globular clusters or dwarf galaxies are expected to be cold substructures , with velocity dispersion @xmath104 @xmath13 . given the errors in @xmath61 , @xmath95 should be larger than 17 @xmath13 . since halo substructures are unlikely to have uniform sizes or velocity dispersions , we ran the group detection algorithm and the monte carlo simulations using combinations of @xmath94 from 1 to 4 kpc in steps of 0.5 kpc and @xmath95 from 20 to 30 @xmath13 in steps of 5 @xmath13 . a group in the rrls data was only considered real when less than 5% of the 5000 monte carlo simulations ( a @xmath105 detection ) produced a group with @xmath106 members within @xmath12 @xmath13 of @xmath107 , the absolute value of the mean velocity of the group of real rrls . the simulations therefore counted random groups near the @xmath72 of the suspected group in the real data and ones near @xmath108 . if a group was detected with different sets of parameters , as happened in most cases , we report here the parameters that gave the lowest probability of random occurrence and/or the maximim number of group members . we restricted our sample to the 33 stars that have @xmath109 kpc since the incompleteness that is caused by the different saturation levels of the ccds of the quest camera @xcite may effect the interpretation of results at closer distances . two significant groups were found among our rrls ( figure [ fig - groups ] ) . as we suspected from the histograms in figures [ fig - vgsr_histo ] and [ fig - bin ] , there is a group with large positive velocity in our data . the maximum number of members ( 8 rrls ) was obtained with @xmath110 kpc and @xmath111 @xmath13 . not one of the 5000 monte carlo simulations produced a similar group . the stars making up this group are spread in @xmath8 from @xmath112 to @xmath113 kpc ( see figure [ fig - groups ] ) . four of the more distant stars in this group ( @xmath114 kpc ) are part of the near side of the `` @xmath115 hr clump '' . the unusually high velocities of these stars and that the fact that some of them are probably part of the `` @xmath115 hr clump '' , unquestionably an over - density in rrls , suggest that they constitute a halo substructure that may be part of the vod ( see [ sec - dis ] ) . the combination of large @xmath72 and a spread in @xmath8 is consistent with a stellar stream that obliquely cuts the line - of - sight ( see fig . 15 in * ? ? the other significant group contains 9 rrls . while it is detected with several choices for @xmath94 and @xmath95 , the values of 1.5 kpc and 25 @xmath13 , respectively , yield the lowest probability of a random occurrence ( only 0.7% of the monte carlo simulations produced similar groups , a @xmath116 detection ) . the relatively small value of @xmath94 of this group indicates that it is more concentrated in space than the high velocity group ( see figure [ fig - groups ] ) . it lies @xmath117 kpc from the sun and has a mean @xmath72 of @xmath118 @xmath13 . this group may be due to a stellar stream that cuts the line - of - sight at nearly a right angle , which would explain the narrow distribution in @xmath8 and the relatively low @xmath72 of the stars . if any of the features described above are real substructures , then they should also appear in the distributions of the other types of stars that make up very old stellar populations . while bhb stars and rrls are , of course , samples of horizontal branch stars that differ primarily in temperature , there are good reasons to suspect that they may not always trace the same stellar populations . the majority of the globular clusters in the outer halo have red hb s , while the metal - poor globular clusters in the inner halo have primarily blue hb s ( the dependence of the second parameter of hb morphology on galactocentric distance ) . consequently , there are large differences in the frequencies of rrls and bhb stars between the clusters in the inner and outer halos @xcite . the majority of the dwarf spheroidal satellite galaxies of the milky way and m31 have red hb s and high frequencies of rrls and low frequencies of bhb stars @xcite . @xcite have suggested that this variation in the rrls / bhb star ratio and the kinematic disconnect between the inner and outer halos ( see [ sec - intro ] ) may be responsible for the kinematic differences that they found between samples of rrls and bhb in the same region of the galactic halo . these observations and the trend in rrls / bhb suggest that the rrls may be a denser tracer of substructure in the outer halo than bhb stars . since a small fraction of outer halo globular clusters and dsph galaxies have blue hb s , each with minority populations of rrls , and since a small population of bhb stars exists in the red hb clusters , some correlation in phase space is expected between these types of stars in the outer halo . on the other hand , an exact correspondence would be surprising . the sample of bhb stars considered here was drawn primarily from the survey by @xcite that is based on photometry and spectroscopy from the sdss . this database of bhb stars is less complete than our rrls sample because the sdss spectroscopy targeted primarily quasars and galaxies . the recent , more complete sample of bhb stars studied by @xcite unfortunately does not overlap with the virgo region . from the @xcite catalog , we selected all bhb stars ( 35 in total ) within the same range of @xmath4 as the rrls sample . in @xmath6 we expanded the limits to @xmath119 . this is still a small range compared with both the width of the region in @xmath4 , and the size of the vod . we also constrained the distance range to @xmath120 kpc . @xcite estimate that their measurements have an average error in velocity of about @xmath121 @xmath13 . it has been established both from observations of globular clusters and theoretical modelling that there is a relation between the luminosity of bhb stars and their temperature in the sense that bluer bhb stars are fainter than red bhb stars ( see for example * ? ? ? when estimating the distances to their bhb sample , @xcite adopted instead the same @xmath122 for all bhb stars on the basis of some observational evidence that they realized was in conflict with theoretical computations . @xcite noted that the adoption of the variation in @xmath122 with effective temperature that is suggested by the computations would lead to smaller distance moduli by an average of @xmath123 mag . @xcite have used the color - magnitude diagrams of globular clusters to derive a relationship between the @xmath124 colors of bhb stars and their @xmath125 values , which is anchored on the same @xmath125 for rrls that we are using . a comparison , after making the necessary transformations between photometric systems , of the @xcite relationship to the one employed by @xcite reveals that the difference in distance modulus is near zero at the blue edge of the instability strip and again small at the highest temperatures , but is @xmath126 mag ( a factor of @xmath127 in distance ) at @xmath128 or @xmath129 . because only a fraction of the sample of bhb stars observed by @xcite _ may _ be effected by distance errors , we have not attempted to recalibrate their distance scale . none of the following comparisons depend critically on an exact match of the bhb and rrls distance scales . in a few of the fields that were observed with the hydra multi - object spectrograph on the wiyn telescope , we identified potential bhb stars from the quest photometry . the three stars that listed in table [ tab - bhb ] appear to be true bhb stars . their @xmath130 and @xmath131 colors , as measured by the sdss , are consistent with the low gravities of bhb stars on the basis of figure 10 in @xcite . > from our spectra , we measured the full widths of the @xmath32 , @xmath40 , and @xmath6 balmer lines at 0.2 below the continuum level , which are well - known diagnostics of surface gravity at the temperatures of bhb stars ( eg . * ? ? ? in addition to the candidate bhb stars , we measured the wiyn spectra of two bhb stars in the globular cluster m3 that are similar in @xmath132 color to the candidates and of the bright a and f type stars that we employed as radial velocity standards with the wiyn observations . the candidate bhb stars have smaller h - line widths than the main - sequence stars and ones that are similar to the bhb stars in m3 . to permit direct comparison with the @xcite sample of bhb stars , we adopted their method to derive the distances to these bhb stars ( see table [ tab - bhb ] ) . in figure [ fig - bin_bhb ] , we have plotted histograms of @xmath72 for the bhb stars from @xcite and table [ tab - bhb ] that lie in the same @xmath8 intervals as the ones in figure [ fig - bin ] . in this diagram , there may be more smearing between the intervals than for the rrls because the absolute magnitudes of the bhb stars are more uncertain ( see above ) . neither of the two features identified in the rrls sample is recognizable in the bhb star data alone , which could be due to the incompleteness of the sample or to low frequencies of bhb stars in these groups . however , several bhb stars have @xmath72 values that coincide with the peaks in the rrls histogram . indeed , when we introduced the combined sample of rrls and bhb stars into the group detection algorithm described above , the same two groups identified in the rrls sample alone appeared . the group with negative velocities now contains 14 members , 5 of which are bhb stars . the new mean velocity is @xmath133 @xmath13 , which we adopt as the velocity of the group . only 5% of the monte carlo simulations produced similar groups . the group with high positive velocities has 10 members ( 2 bhb stars ) . similar groups were obtained in only 2 out of 5000 simulations ( 0.04% ) . the combined sample of rrls and bhb stars also contains a third , small but nonetheless significant , group having a mean radial velocity of @xmath134 @xmath13 . the group , which was detected with @xmath135 kpc and @xmath111 @xmath13 , contains only 3 members , all of which are bhb stars located in the very narrow intervals of @xmath136 to @xmath137 kpc and @xmath138 to @xmath139 . the monte carlo simulations found simialr groups in @xmath140 of the cases ( a @xmath141 detection ) . this group may be related to main sequence stars with similar velocities ( see below ) and with 2 rrls at essentially the same distance and velocity but farther west ( @xmath142 ) . since this group lies near the faint limit of our sample , it may be the near side of a larger substructure . the recent investigation of the vod region by @xcite has provided radial velocity measurements for many stars with the colors of f type stars in two segue fields , at ( @xmath143 ) = ( @xmath144 ) and ( @xmath145 ) , one of which overlaps with the region we are considering here ( see figure [ fig - zone ] ) . @xcite make the very reasonable assumption that the vast majority of these stars are probably near the main - sequence turnoff in the color - magnitude diagram . because such stars have a significant spread in absolute magnitude , their distances are much less precisely known than either the rrls or the bhb stars . the big advantage of using these stars is that they are much more numerous than either rrls or bhb stars . below we examine their data for evidence for the major groups described above . [ [ group - at - v_gsr215-rm - kms-1 . ] ] group at @xmath146 @xmath13 . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + this group is probably the most remarkable feature seen in our data since very few stars with such high velocities are expected in a sample of halo stars ( 1 or 2 in our sample of rrls ) . the 10 rrls and bhb stars in this group have a mean velocity of @xmath147 @xmath13 , and a standard deviation of @xmath148 @xmath13 , which is not much larger than the typical error in our radial velocities ( @xmath149 @xmath13 ) . as noted above , this group appears to be spread along the line of sight , and it may be significant that in the sample of rrls studied by @xcite , there are 3 stars with @xmath72 values consistent with membership in this group that have @xmath8 between 16 and 17 kpc . in their discussion of the f - type stars , @xcite briefly mention an excess of high velocity stars , which we suspect is related to this group of rrls and bhb stars . in their figure 9 , @xcite present for each of their fields histograms of @xmath72 for the f stars in two ranges in @xmath150 magnitudes . in the histogram for the brighter group , which according to the distance scale , then the distance uncertainty may be as large as 40% ] of @xcite corresponds to 11 to 14 kpc , there are 8 f type stars with @xmath72 consistent with assignment to our high velocity group . it is probably significant that all 8 of these stars are found in the more northern of the two fields ( see figure [ fig - zone ] ) , the one that overlaps with the region studied here ( 2 stars of even higher @xmath72 are present in this field as well ) . since neither one of the two f - star fields has any stars with @xmath151 @xmath13 , there is an asymmetry toward large positive velocities in the same region of the sky where we find one in the rrls sample . among the fainter f type stars , which lie at distances between 14 and 18 kpc , there are only 3 stars in the 2 fields combined with @xmath72 within the range of this group . we have been able to determine metallicities for 6 of the 8 rrls in our sample that appear to be members of this group . they have a mean @xmath152 } \rangle = -1.55 $ ] , with a standard deviation of 0.15 dex , which is of the same size as the uncertainty in our [ fe / h ] measurements . this narrow distribution is much more consistent with the debris from a destroyed globular cluster than from a dwarf galaxy . two of the 3 high - velocity rrls measured by @xcite have metallicities that are similar to the metallicities of the brighter rrls . the other one has a much lower ( [ fe / h]@xmath153 ) , which may be a sign that it is unrelated . we are observing more rrls in the range @xmath154 kpc , with the expectation that more velocity and metallicity measurements will reveal the origin of this substructure . [ [ group - at - v_gsr-49-rm - kms-1 . ] ] group at @xmath155 @xmath13 . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + this is the most populated group that we find in our data , containing 9 rrls and 5 bhb stars with velocities in the range @xmath156 @xmath13 . the mean velocity of this group is @xmath157 @xmath13 , with a standard deviation of @xmath158 @xmath13 . the contamination from unrelated halo stars is expected to be relatively high at the low velocity of this group , and assuming the normal distribution in figure [ fig - vgsr_histo ] for such field stars , we estimate that only about 50% of the stars in this velocity interval are probably members of a stream . thus , our estimates for both the mean velocity and the velocity dispersion are likely to be skewed by this contamination . the 9 rrls in this group have @xmath159 } \rangle = -1.72 $ ] , with a standard deviation of 0.28 dex . there is not a strong peak in the metallicity distribution that would suggest that the large dispersion is caused by contamination by non - members of an otherwise narrow distribution . @xcite found a significant peak at @xmath160 @xmath13 in the velocities of the f - type stars , which considering the uncertainties is close to the peak that we find . while this suggests membership in the same halo substructure , it is not altogether clear that they overlap spatially . this feature is most clearly evident in the southern of the two fields investigated by @xcite , the one that is outside the region investigated here , and among the bluer and brighter sample of f - type stars ( @xmath161 kpc if they have @xmath162 ) . on the other hand , the rrls in this group have @xmath163 kpc , which place them in the foreground of the f - type stars , unless the f - type stars are less luminous than suspected . there is better spatial coincidence between the f - type stars and a group of fainter rrls that were measured by @xcite , which have similar @xmath72 . the relationships between the f - type stars , the rrls in @xcite , a sample of m giants , and the tidal stream from the sgr dwarf galaxy are discussed at length by @xcite , who suspect that this feature may be a new halo substructure that is unrelated to sgr . the feature that we have identified appears to be unrelated to sgr as well ( see below ) . more observations are required to clarify its relationship to the feature described by @xcite . [ [ group - at - v_gsr-171-rm - kms-1 . ] ] group at @xmath164 @xmath13 . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the 3 bhb stars in the group have basically the same values of @xmath8 , 11.2 kpc , but given the distance uncertainty a spread of @xmath165 kpc can not be excluded . @xcite identify a peak in their f star data at @xmath166 @xmath13 , and in this case , there are good reasons to believe that the features are the same . a peak is seen in both of their fields , although it is more prominent in the southern field , the one that is outside our region , than in the northern . it is only seen in the brighter f stars that have @xmath167 . according to the distance scale of @xcite these stars lie at @xmath8 between 11 and 14 kpc , which agrees well with the distances of the bhb stars . the f star data show a very prominent peak at @xmath168 @xmath13 in both fields , and @xcite consider this to be the velocity of s297 + 63 - 20.5 feature . on the basis of the spatial coincidence between s297 + 63 - 20.5 and the vss and the near coincidence in @xmath72 , they conclude that these features are probably the same or physically associated . however , @xcite note that the @xmath169 @xmath13 offset in @xmath72 between s297 + 63 - 20.5 and the vss requires explanation . the peak at @xmath168 @xmath13 is seen in the faintest and the reddest group of f stars , which have @xmath150 between 20 and 20.5 . @xcite note , however , that the most of these stars actually have @xmath170 , which corresponds to @xmath171 kpc on their distance scale . since the core of the vss is located at about 19 kpc @xcite , it is possible that @xmath169 @xmath13 offset is due to a gradient in @xmath72 with @xmath8 . some evidence for this is provided by the data for the brighter and bluer f star data ( see figure 10 in @xcite ) , which has a prominent peak at @xmath172 @xmath13 . the histograms in figures [ fig - bin ] and [ fig - bin_bhb ] indicate there are 5 stars ( 2 rrls and 3 bhb stars ) having a mean velocity of @xmath173 @xmath13 which , within uncertainties , is identical to the peak seen in the bright f star data . all 5 rrls and bhb stars are located near the far distance limit of our sample at @xmath174 kpc and within the near boundary of the `` 12.4 hr clump '' of rrls . the 5 stars are indeed identified as a group by our algorithm ( with parameters @xmath175 kpc , @xmath176 @xmath13 ) , but by itself this group has only low significance because @xmath177 of the monte carlo simulations produce similar groups . while a chance coincidence can not be ruled out , the similarity in @xmath178 between this group and the peak in the bright f star data suggests that they are related . the group of rrls and bhb stars provides a hint that the vss may extend to distances as short as @xmath179 kpc . a gradient in @xmath72 , with @xmath72 becoming less positive with increasing @xmath8 , is expected of a stellar stream that obliquely cuts the line of sight ( see * ? ? ? it is also possible , however , that observational error and sample selection may also explain the @xmath72 offset . we are exploring the possible extension of the vss and the possibility of a gradient by measuring more quest rrls at distances between 13 to 18 kpc ( duffau et al . 2008 , in preparation ) . soon after the first detections of over - densities of stars in the direction of virgo , several authors ( eg . * ? ? ? * ; * ? ? ? * ) speculated that they may be debris from the disruption of the sgr dsph galaxy . even then there was some evidence to contrary , for @xcite had noted that the color of the turnoff in s297 + 63 - 20.5 was redder than the color of the turnoff in the sgr streams . @xcite later argued that the sgr streams were not responsible for the vss or the majority of the other stars in their sample because models of the streams produced very few stars at the observed positions and the predicted values of @xmath72 did not match the ones observed . the realization by @xcite that the vod has huge dimensions on the sky motivated @xcite to reexamine the sgr connection . through detailed modeling of the streams , they showed that ones that assume an oblate shape for the gravitational potential of the galaxy can explain the vod as the leading stream from sgr as it plunges toward the sun and crosses the galactic plane in the solar neighborhood . this model does not , however , provide a good match to our observations , those of @xcite , and more generally the 12.4 hr clump of rrls . in the left - hand panels of figure [ fig - sgr ] , we have plotted the rrls in the first band of the quest survey that lie within the @xmath4 range of our sample . closed symbols depict the rrls that have measured velocities from @xcite , @xcite , and the present study . these three datasets cover different range of distances , @xmath180 , @xmath181 and @xmath182 kpc respectively . the small crosses are the particles in the stream models of @xcite under the assumptions that the potential is oblate , spherical , or prolate . the panels on the right compare the model predictions for @xmath72 as a function of @xmath8 with the observed velocities . the quest survey detects the sgr leading arm at @xmath183 kpc in the range @xmath184 ( see figure [ fig - quest ] and vivas et al . 2005 ) and the 12.4 hr clump in the ranges @xmath185 kpc and @xmath186 ( see fig . 12 in * ? ? ? note that there is not a blending together of these features ( see figure [ fig - quest ] ) , as one would expect if they are parts of the same stream , but instead they are separated by a region of low density . this is not a selection effect introduced by the observing technique ( see * ? ? ? if the leading arm crosses the galactic plane near the sun , then the stream stars at small @xmath8 will have large , negative values of @xmath72 because they have fallen from @xmath31 kpc on orbits that are nearly aligned with the line - of - sight from the sun . this prediction of large negative velocity at small @xmath8 and the continuity of the stream are two characteristics of this picture that should not be model dependent . @xcite note that their models reproduce the models of @xcite , which because they are readily available , we will use here to illustrate the salient points of this picture . each of the @xcite models matches reasonably well the positions and the velocities of the rrls at @xmath31 kpc . however , both the oblate and the spherical models predict smoothly varying densities of rrls over wide ranges in @xmath8 , which are in conflict with the observed region of relatively low density of rrls between 20 and 40 kpc ( see also * ? ? ? these same models predict that @xmath72 becomes more negative as @xmath8 decreases and reaches quite large @xmath187 at @xmath188 kpc . only one rrls in our sample and one in the sample observed by @xcite have values of @xmath72 that are consistent with these models . also , none of the bhb stars in the region have highly negative velocities ( see figure [ fig - bin_bhb ] ) . in the prolate model , the leading stream crosses the galactic plane far from the sun , which explains why it predicts relatively few stars in the diagrams at @xmath188 kpc . we conclude on the basis of these comparisons that none of the halo substructures that we have detected are probably due to the leading arm . moreover , based on the current sgr models , our observations imply that the shape of the galactic potential can be neither oblate nor spherical . @xcite note that in their oblate model the trailing stream of sgr crosses through the region of the vss and coincides with the vss in @xmath72 ( @xmath189 km / s ) . this is a plausible explanation for the vss , if it is still viable after the oblate model is reconciled with the observations presented here and with other recent observations which indicate that the leading arm does not cross the galactic plane in the solar neighborhood @xcite . kinematical information is crucial to detect substructures in regions where no obvious spatial overdensities are present . taken together , our rrls sample , the bhb star sample of @xcite , and the f star sample of @xcite paint a consistent picture that several separate halo substructures exist at @xmath190 kpc in the part of the vod that is covered by the quest survey . the rrls sample of @xcite and the same bhb and f star samples show that some of these substructures extend to larger @xmath8 . none of these substructures seem to be related with the sgr leading tail . the analysis of the sdss photometric data by @xcite make a compelling case that the region of highest stellar density in s297 + 63 - 20.5 lies at @xmath191 kpc . this is identical to within the errors of the region of highest density in the 12.4 hr clump of quest rrls , which is where @xcite discovered the vss in velocity space . the near coincidence in @xmath178 between the vss and the largest peak in @xmath72 that @xcite observed in s297 + 63 - 20.5 provides strong evidence that they are the same . as noted by @xcite , the small offset ( @xmath192 @xmath13 ) could be due to measurement error and/or by the small sizes of the samples of spectroscopically observed stars . it is also possible that it is due to a gradient in @xmath72 with @xmath8 ( see above ) . @xcite found the same velocity peak in their two spectroscopic fields , which indicates that the s297 + 63 - 20.5/ vss has considerable angular extent on the sky . other evidence for this is provided by the recent rrls surveys by @xcite and @xcite . @xcite found 2 clumps of rrls south of the vod region ( see figure [ fig - zone ] ) at coordinates @xmath193 and @xmath194 and distances 16 and 19 kpc respectively . velocity information is not yet available for these clumps . @xcite identified an excess of candidate rrls having @xmath195 @xmath13 in the north galactic cap ( @xmath196 ) . the results presented here indicate that the s297 + 63 - 20.5/vss feature may also extend along the line of sight to distances as close as @xmath197 kpc . we conclude that the vod contains at least one relatively large structure , s297 + 63 - 20.5 /vss , which is intersperced with smaller halo substructures that are unrelated to s297 + 63 - 20.5/vss or to each other . while these smaller structures might be peculiar to the vod region , it is more likely that they are common in the outer halo . they are better documented in the vod direction simply because the presence of s297 + 63 - 20.5/ vss has drawn attention to this region . in the case of our sample of rrls , the spatial distribution of the stars provided only very marginal evidence of any substructure @xcite . the picture changed dramatically when we used spatial and velocity data together . there is already strong evidence from the spatial distribution of stars in the sdss that much , if not all , of the outer halo was accreted from the destroyed of satellite galaxies @xcite . our investigation of the vod region suggests that the addition of velocity data will make the case even stronger . this research was supported by the national science foundation under grants ast 00 - 98428 and ast 05 - 07364 . the availability of the smarts 1.5 m telescope and its service observing mode made this project feasible . we thank charles bailyn for his efforts to setup the smarts consortium and fred walter and suzanne tourtellotte for their assistance with smarts operations and data handling . we are indebted to the service mode observers sergio gonzalez , alberto pasten , claudio aguilera and alberto miranda for their fine operation of the telescope and the spectrograph . we also thank the anonymous referee for helpful suggestions to the manuscript . , r. , vivas , a. k. , gallart , c. , & winnick , r. 2004 , in astronomical society of the pacific conference series , vol . 327 , satellites and tidal streams , ed . f. prada , d. martinez delgado , & t. j. mahoney , 92 to @xmath201 ( from * ? ? ? the box encloses the bright subsample ( @xmath202 ) toward virgo for which we obtained spectroscopy . the @xmath200 symbols indicate the 8 rrls from the second quest catalog with available spectroscopy that are included in our analysis . the large spatial substructures found among the fainter rrls ( see * ? ? ? * ) are indicated : the sgr stream , the globular cluster pal 5 and its tails , and the `` @xmath203 clump '' , which contains the vss . we have indicated with solid circles the rrls having velocities consistent with the vss @xcite . ] . the two significant groups found in our data are indicated by solid symbols , one of them have stars with very high positive velocities and the other one has stars with velocities around @xmath204 @xmath13 . ] lccccccl eso / la silla & 1.5 m & b&c & 600 & 3300- 5500 & 3.1 & 1/1 & 2001 jun 17 + kpno & wiyn & hydra & 600 & 7100 - 10000 & 2.9 & 5/3 & 2003 mar 11 + ctio & smarts 1.5 m & r - c & 600 & 3532 - 5300 & 4.3 & 42/28 & 2003 may 1 - 9 + ctio & smarts 1.5 m & r - c & 600 & 3532 - 5300 & 4.3 & 24/14 & 2004 feb - jun + ctio & smarts 1.5 m & r - c & 600 & 3532 - 5300 & 4.3 & 5/4 & 2006 jun + ctio & smarts 1.5 m & r - c & 600 & 3532 - 5300 & 4.3 & 4/2 & 2007 feb - apr + kpno & wiyn & hydra & 400 & 3500 - 6200 & 7.1 & 9/5 & 2007 feb 20 - 21 + kpno & wiyn & hydra & 600 & 7100 - 10000 & 2.9 & 3/3 & 2007 apr 27 , may 2 + [ tab - telescopes ] cccccrrc 201 & 2452711.7580 & 3600 & wiyn - r & 0.76 & 58 & 5 & + 201 & 2452765.6208 & 1200 & smarts & 0.02 & 44 & 16 & -1.95 + 203 & 2452765.6385 & 1200 & smarts & 0.02 & 13 & 16 & -1.40 + 212 & 2452078.4665 & 900 & eso & 0.56 & 88 & 11 & -2.11 + 212 & 2452760.6437 & 300 & smarts & 0.58 & 109 & 16 & -1.99 + [ tab - observations ] ccccccccccccccc 151 & 169.625160 & -0.764498 & 14.78 & c & 0.29370 & 2451611.7002 & 7.0 & 2 & 2 & 90 & 28 & 28 & -35 & -1.32 + 157 & 171.487530 & -0.161508 & 14.39 & ab & 0.70840 & 2451571.8154 & 5.9 & 2 & 2 & 142 & 22 & 22 & 22 & -1.67 + 160 & 172.957305 & -2.240574 & 15.43 & ab & 0.67619 & 2451582.7058 & 9.4 & 2 & 1 & 217 & & 20 & 93 & -2.52 + 161 & 173.844930 & -0.895030 & 15.40 & ab & 0.60699 & 2451255.6642 & 9.3 & 2 & 2 & 18 & 7 & 15 & -100 & -1.71 + [ tab - velocities ]	using a sample of 43 bright ( @xmath0 , distance @xmath1 kpc ) rr lyrae stars ( rrls ) from the quest survey with spectroscopic radial velocities and metallicities , we find that several separate halo substructures contribute to the virgo overdensity ( vod ) . while there is little evidence for halo substructure in the spatial distribution of these stars , their distribution in radial velocity reveals two moving groups . these results are reinforced when the sample is combined with a sample of blue horizontal branch stars that were identified in the sdss , and the combined sample provides evidence for one additional moving group . these groups correspond to peaks in the radial velocity distribution of a sample of f type main - sequence stars that was recently observed in the same directon by segue , although in one case the rrls and f star groups may not lie at the same distance . one of the new substructures has a very narrow range in metallicity , which is more consistent with it being the debris from a destroyed globular cluster than from a dwarf galaxy . a small concentration of stars have radial velocities that are similar to the virgo stellar stream ( vss ) that was identified previously in a fainter sample of rrls . our results suggest that this feature extends to distances as short as @xmath2 kpc from its previous detection at @xmath3 kpc . none of the new groups and only one star in the sample have velocities that are consistent with membership in the leading tidal stream from the sagittarius dwarf spheroidal galaxy , which some authors have suggested is the origin of the vod .
the wavefunction of a single photon travelling through the photonic chip can be described in the waveguide basis @xmath112 where @xmath113 denotes the presence of a photon in waveguide @xmath114 . the photon travels through a series of phaseshifters @xmath115 and directional couplers @xmath116 . the action of each linear element on the wavefunction can be described by a 6 x 6 matrix such that the wavefunction after the element is @xmath117 a phaseshifter applying the phase shift @xmath118 to the photon in waveguide @xmath119 is described by the matrix @xmath120,\ ] ] where @xmath20 and @xmath19 label the waveguides . for example a phaseshifter on waveguide 2 is given by @xmath121 which is a unity matrix , where the 2nd element on the diagonal is replaced by a phase factor . a directional coupler between waveguide @xmath119 and @xmath122 with reflectivity @xmath123 is described by @xmath124\nonumber \\ & & + \left(1-\delta_{ak}-\delta_{bk}\right)\delta_{kl}.\end{aligned}\ ] ] as an example a directional coupler with reflectivity @xmath123 between waveguide 2 and waveguide 3 is given by @xmath125 the unitary of the chip @xmath26 has a rather complicated form and we will therefore present it as the product of its elementary building blocks . it can be written in terms of the preparation stage unitary @xmath126 and the unitary describing the selection of the measurement bases @xmath127 as @xmath128 where the preparation stage unitary can be broken down further into @xmath129 with the unitary of the central part given by @xmath130 where the reflectivities @xmath131 ( design and measured ) can be found in tab [ tab1 ] . the two other blocks are given by @xmath132 @xmath133 in the same way the unitary describing the selection of the measurement bases can be written as @xmath134 with @xmath135 @xmath136 this fully defines the single photon unitary of the photonic chip under consideration . once the unitary is known we can proceed to calculate the probabilities for coincidence counts at the detectors . we will assume that the source produces photon with a given frequency distribution , so that a photon in wave guide @xmath85 is described by @xmath137 where @xmath138 is the vacuum state , @xmath139 a creation operator for a photon with frequency @xmath140 in waveguide @xmath85 and @xmath141 a normalised amplitude for a frequency so that @xmath142 . with a photonic circuit described by the single photon unitary @xmath26 and an initial state with two photons entering the waveguides @xmath143 and @xmath144 , @xmath145 , we can write the wave function at the exit of the circuit as @xmath146 where we defined the two photon state @xmath147 the coincidence counts at detectors @xmath20 and @xmath19 , @xmath23 , are then given by @xmath148 where the number operators are given by @xmath149 after a lengthy but straightforward calculation , contracting pairs of creation and annihilation operators , using the commutation relation @xmath150=\delta_{ij}\delta(\omega_{1}-\omega_{2})\ ] ] we obtain @xmath151 with @xmath152 note that we have here selected the frequency distribution to distinguish the photons but a similar argument can be made for other degrees of freedom , e.g. polarisation . in our experiments we are interested in the the probability of a certain coincidence , given that we insert a photon into each of the waveguides 2 and 4 . we obtain the probability of coincidence as @xmath153 @xmath154 with the normalisation constant given by @xmath155 which is exactly the form of eq . ( [ coincidenceprob ] ) . we now show that @xmath58 can be directly obtained from a hong - ou - mandel type experiment . in such a measurement two photons are directed to two different ports of a directional coupler with reflectivity @xmath123=1/2 . the unitary is given by @xmath156 what is measured is the number of coincidence counts relative to the number of counts obtained for completely distinguishable photons . for a beam splitter with reflectivity @xmath123 we obtain @xmath157 normalising this to the number of coincidence counts for completely distinguishable photons we get @xmath158 or expanding around @xmath159 to second order in @xmath160 we obtain @xmath161 this shows that we can directly read off the function @xmath58 from a hong - ou - mandel type experiment and that the error induced by using a non - ideal beam splitter is quadratic in the deviation . if we want to consider the photonic quantum circuit in the light of quantum computation we have to assign the meaning of qubits to certain combinations of photons . we do this in the following way : waveguides 2 and 3 encode one qubit in dual rail encoding , while waveguides 4 and 5 encode the second one . a photon present in waveguide 2 and the second photon in waveguide 4 maps onto the logical two qubit state @xmath162 . similarly we make the mappings @xmath163 , @xmath164 and @xmath165 . given the full wave function on the chip and the fact that the detectors only register the presence of a photon , but do not distinguish parameters like polarisation or wavelength , we can construct a reduced density matrix for the two qubit subspace as @xmath166 where the pairs @xmath167 and @xmath168 are taken from @xmath169 the coefficients are given by @xmath170 with @xmath171 given by eq . ( [ largepsi ] ) and @xmath172 defined in eq . ( [ twophotonstate ] ) and the partial trace running over the frequency degree of freedom . in a similar calculation to the previous section and assuming that single photons are inserted in waveguide 2 and 4 we obtain @xmath173 with @xmath174 defined in eq . ( [ pdist ] ) . reinserting this in the equation for the density matrix eq . ( [ dmatr ] ) we see that we can define @xmath175 and	integrated quantum photonic circuits are becoming increasingly complex . accurate calibration of device parameters and detailed characterisation of the prepared quantum states are critically important for future progress . here we report on an effective experimental calibration method based on bayesian updating and markov chain monte carlo integration . we use this calibration technique to characterize a two qubit chip and extract the reflectivities of its directional couplers . an average quantum state tomography fidelity of [email protected]@xmath1 against the four bell states is achieved . furthermore , comparing the measured density matrices against a model using the non - ideal device parameters derived from the calibration we achieve an average fidelity of [email protected]@xmath1 . this pinpoints non - ideality of chip parameters as a major factor in the decrease of bell state fidelity . we also perform quantum state tomography for bell states while continuously varying photon distinguishability and find excellent agreement with theory . a large variety of quantum systems have been studied in recent years for use in quantum information processing @xcite . among the possible implementations , quantum information processing based on photons stands out due to its high stability , wide availability and ease of manipulation @xcite . recent developments in integrated photonics have been promising from the point of view of future large scale on - chip quantum information processing @xcite . moreover , integrated reconfigurable quantum photonic circuits have shown great potential for generic quantum operations @xcite . for instance , it has been shown that the generation of arbitrary two - qubit states and the corresponding state tomography can be realised on a chip with high fidelity @xcite . however , with increasing device complexity the accurate calibration of quantum devices becomes a crucially important task . in this paper we study a reconfigurable two - qubit quantum photonic device designed to create maximally entangled states and to perform quantum state tomography on them . in particular we focus on statistically rigorous calibration and tomography which allow us to reach a very high fidelity between theoretically expected and measured states . the paper is organised as follows . we first describe the experimental arrangement and the device design . we continue by detailing the statistical calibration procedure and the used theoretical model that takes into account the finite quantum interference . we also explain how the state tomography is performed . we then proceed to the benchmarking results for states ranging from fully mixed to nearly maximally entangled by adjusting photon delay . we conclude by discussing the results . the photonic circuit investigated here was fabricated using silica - on - silicon technology @xcite . figure [ fig1 ] shows a schematic design of the device . a qubit is encoded in the amplitude and phase of a single photon travelling on a pair of waveguides ( path encoding ) @xcite . for realising a two - qubit state two identical photons and four waveguides are necessary . for instance one photon at each of the inputs 2 and 4 corresponds to the two qubit state @xmath2 . the chip can be viewed as composed of three parts : the first part on the left prepares arbitrary single qubit states ( see the pink p / m1 and blue p / m2 blocks in fig . [ fig1 ] ) . the central part ( see the yellow block c in fig . [ fig1 ] ) is responsible for the quantum entanglement owing to the probabilistic controlled - not gate ( cnot ) @xcite . the blocks on the right hand side are mirror images of the preparation blocks on the left and are used to choose the basis for projective measurements on the two single qubits . each block consists of a number of directional couplers ( dc ) and voltage - controlled thermal variable phase shifters @xcite . the experimental setup that we used is similar to that of ref . @xcite , the only difference being the reconfigurable photonic chip . the chip was mounted on a chip holder and butt coupled with optical fibre v - groove arrays at the input and output . photon pairs were generated by a type - i spontaneous parametric down - conversion source , pumped using a 50 mw 405 nm laser . the 810 nm photons were filtered with 2 nm bandpass filters and collected into polarisation maintaining fibres with aspheric lenses and then directed to the chip . the detection of photons was done by single photon avalanche diodes connected to the output fibre array with an efficiency of about 50@xmath1 . all photon arrival times were recorded by a counting card at a time resolution of about 165 ps . two photon coincidences could then be detected between every pair of output waveguides . the relative delay of the response time between all the detectors has been calibrated and deducted before counting coincidences . a time window of 1.5 ns was employed to count coincidence in the experiment . this width of the window is required because of statistic variation of detector response time . the random two photon coincidence rate for this time window was calculated in the experiments and was always less than 0.4% . the generation , manipulation and measurement of entanglement and single qubit mixed states has been demonstrated with a linear photonic two - qubit chip of the same design @xcite as in the present paper . in those experiments an average quantum state tomography fidelity of [email protected]@xmath1 was achieved for the four maximally entangled bell states . however , from experiments performed to date , it is unclear what mechanisms cause the decrease from ideal fidelity . possible causes include distinguishability of the input photon pair , inaccurate phase shifters , non ideal reflectivities of the on - chip directional couplers , variations in the output coupling and photon detection efficiencies . in order to make further improvement to the fidelity of chip operation it is important to carefully characterise all these . we focus here on the chip parameters and distinguishability . schematic design of the integrated photonic device . the reconfigurable phase shifters are labelled by @xmath3 and the waveguide directional couplers by dc1 - 13 . all directional couplers have a design reflectivity of @xmath4 except the ones marked by a dot ( reflectivity @xmath5 ) . the shaded blocks indicate different functions within the circuit . the pink p / m1 and blue blocks p / m2 perform single qubit rotations on the target qubit and the control qubit , respectively . the yellow region c performs a cnot operation . the arrows on the input side indicate the input waveguides for the photon pair arriving from the source . we adopt a convention such that the control qubit states @xmath6 and @xmath7 correspond to the output channel pairs 5 and 4 , respectively , while the target qubit states @xmath6 and @xmath7 correspond to the output channel pairs 3 and 2 , respectively.,scaledwidth=90.0% ] before using the bayesian technique to calibrate the the chip ( passive ) parameters accurate mapping of the phase shift as a function of the applied voltage is required . the eight resistor - based variable thermo - optic phase shifters dependencies were measured individually with single photons @xcite . this was done using interference by sending single photons repeatedly to the waveguides 2 or 4 and then counting the number of output photons within a fixed time window as a function of the heater voltage . all the eight calibrations were done while the remaining heaters were driven at a medium power level to mimic the conditions during a typical experiment . with all the eight phase shifters calibrated , the reflectivities of the 13 on - chip directional couplers @xmath8 could be determined using bayesian inference @xcite . a statistical model for the set of variables @xmath9 given the set of parameters @xmath10 is described by the conditional probability @xmath11 . this model can be converted to the distribution of the parameters @xmath12 given a set of observations @xmath9 using the bayes rule @xmath13 where @xmath14 is the prior distribution containing the initial assumptions and @xmath12 is the posterior distribution . in the present context @xmath9 is the set of observed photon coincidences for different configurations whereas @xmath10 is the set of device parameters . the rule can be used iteratively for updating the distribution : the latest posterior can be used as a prior for new data . the method makes the dependence on underlying assumptions transparent and provides a general model that can be applied as long as one is able to write down an appropriate statistical model for the experiment . in the present study the required likelihood function is for two - photon coincidences , given the parameters of the photonic circuit . evaluating the likelihood function exactly is hard and we use a markov chain monte carlo ( mcmc ) @xcite method to draw samples from the posterior . to generate the experimental data for the bayesian inference task we pick 1000 sets of 8 random phases uniformly generated between 0 and 2@xmath15 and bias the phase shifters accordingly . for each of the 1000 phase settings @xmath16 , @xmath17 , identical photons were sent to waveguides 2 and 4 while we recorded the number of coincidences @xmath18 between all the pairs @xmath19 and @xmath20 of the 6 output channels . that is , we recorded the frequency of the 15 different coincidence events for 1000 randomly chosen phase settings . we denote this data set collectively as @xmath21 and the number of coincidences observed for a given phase setting @xmath16 is indicated by @xmath22 . for two channels @xmath23 one can express the expected probabilities for the observed coincidences compactly as @xmath24.\nonumber \end{aligned}\ ] ] here @xmath25 is a parameter describing the probability that the two photons are distinguishable . @xmath26 is the underlying @xmath27 unitary describing the single - photon behaviour of the chip . @xmath28 can be obtained in a straightforward way by combining the effect of the 13 directional couplers and 8 phase shifters ( see [ appendixa ] ) . the probability @xmath29 thus depends on all the chip parameters of interest . the normalisation factor @xmath30 is obtained by summing over all the events that our detection scheme can detect . this is needed since the events corresponding to two photons in the same channel are not measured . for a derivation of the coincidence probabilities starting from the device unitaries see [ appendixb ] . the model for the coincidence probability can be interpreted as a statistical mixture of ideal quantum interference behaviour and distinguishable behaviour . to see how this model can be used to obtain the unknown parameters , let us consider the probability of observing the set of coincidences @xmath21 given the parameters @xmath31 . for each of the experiments @xmath32 the probability of observing a number of coincidences @xmath33 is given by the multinomial distribution @xmath34 we can therefore write the total probability as the product of 1000 multinomial distributions ( for a given total number of events @xmath35 for each experiment @xmath32 ) @xmath36 this function should be considered to be the conditional distribution given that @xmath35 events have occurred for each @xmath32 . although @xmath35 is in principle a stochastic variable , it could just as well be selected by collecting precisely that amount of data . when the @xmath37 are fixed , this function of @xmath38 is called the likelihood function . however , we were interested in the distribution of the parameters given the observations , and not vice versa . we therefore used the bayes theorem to write the so - called posterior distribution for @xmath39 as @xmath40 here the normalisation factor can be in principle obtained by integrating over @xmath39 but in practice it is hard to do without solving for the distribution . here we took the prior as a constant but set it to zero in unphysical regions of the parameter space . it is very difficult to directly evaluate @xmath41 . instead , we used the markov chain monte carlo metropolis - hastings algorithm to perform a random walk in the parameter space @xmath39 . this method works by starting from a random point in parameter space and picking trial points randomly in a symmetric way . a trial point can be accepted or rejected depending on how its probability compares to the probability of the previous point . the new point is always accepted if it is more probable . if the new point is less probable it is accepted with the probability corresponding to the ratio of probabilities . here pseudo random numbers are used . note that using ratios avoids calculating the normalisation factor . numerically it is actually much more accurate and stable to perform the comparisons with logarithms of the pseudo random numbers and logarithms of the probabilities @xmath42 . in our case this approach avoided the need of multiplying 15000 below unity numbers before comparison . it also eliminated the factorial term which is a constant as a function of @xmath39 . the resulting data was stored and owing to so - called detailed balance and ergodicity , the walk sampled the distribution as if the points were drawn from it . one could then calculate e.g. the moments or histograms from the data . for the data shown below we let the system initialise for 30000 steps and we then sampled for 200000 steps . the number of steps was chosen empirically . trials were performed by adding a normally distributed number ( standard deviation was chosen to be 0.5% of design / expected value ) to a randomly picked parameter . without optimising the code the sampling takes about the same time as the experiment using a laptop running matlab ( overnight ) . the resulting expectation values and standard deviations of the reflectivities of the directional couplers 1 - 13 are shown in table [ tab1 ] . the differences between fitted and designed values were expected to be less than 5@xmath1 for the process employed . the slightly larger observed variation indicates that the fabrication process was not fully within specifications . in addition , we found the probability of two - photon distinguishability ( which equals one minus visibility of two photon interference ) to be @xmath43 . this probability of distinguishability takes into account all contributions that might deteriorate the visibility of two photon interference such as non - identical spectra and polarisation of the photon source as well as non - uniform refractive index or birefringence in the waveguides . c > m2.3cm<>m4.3cm < [email protected] & design values & fitted values & + @xmath44 & 0.3333 & [email protected] & -2 & 3% + @xmath45 & 0.5000 & [email protected] & 3 & 7% + @xmath46 & 0.5000 & [email protected] & 1 & 3% + @xmath47 & 0.5000 & [email protected] & -10 & 1% + @xmath48 & 0.5000 & [email protected] & -9 & 5% + @xmath49 & 0.5000 & [email protected] & 7 & 5% + @xmath50 & 0.5000 & [email protected] & 12 & 7% + @xmath51 & 0.3333 & [email protected] & -4 & 7% + @xmath52 & 0.5000 & [email protected] & 7 & 6% + @xmath53 & 0.5000 & [email protected] & 0 & 2% + @xmath54 & 0.5000 & [email protected] & 4 & 1% + @xmath55 & 0.5000 & [email protected] & 15 & 2% + @xmath56 & 0.3333 & [email protected] & -11 & 0% + to confirm the reliability of the fitted results of both the coupler reflectivities and distinguishability of the photons , two photon interference experiments were carried out over various mach zehnder interferometers on the chip . figure [ fig2 ] shows a hong - ou - mandel dip @xcite , measured over the branch on the top right corner of the chip . the hong ou mandel dip was obtained by inserting photons in waveguides 1 and 4 and counting the coincidences at waveguides 2 and 3 . the phase shifters 3 and 4 on the preparation side were adjusted so that the first interference takes place at directional coupler 5 , while phase shifters 5 and 6 were tuned so that the effective reflectivity of the whole branch was as close to 50% as possible ( estimated 52% ) . the visibility of the dip was measured to be @xmath57 . another hong ou mandel dip measurement over the central directional coupler 8 yielded a visibility of [email protected]@xmath1 . this is about 3.1@xmath1 lower than the value ( 76.21@xmath1 ) to be expected from reflectivity of 0.3175 for coupler 8 . both of the hong ou mandel scans thus resulted in about 3 - 4@xmath1 imperfection which is in agreement with our markov chain monte carlo - based characterisation . hong ou mandel dip scan of an on - chip mach zehnder interferometer whose reflectivity is set to 0.5 . the probability of distinguishability @xmath58 can be obtained from a fit to the measurement of the two photon coincidence counts . the maximum number of coincidence counts is identified with @xmath59 . , scaledwidth=90.0% ] having carefully calibrated the integrated quantum photonic chip parameters we then performed a demanding benchmark experiment . the generation and characterization of maximally entangled bell states offered an ideal test case for this purpose . using inputs 2 and 4 we prepared the initial state @xmath2 and drove the input side phase shifters 1 - 4 to prepare pure product states . the corresponding bell states @xmath60 , @xmath61 , @xmath62 and @xmath63 could then be produced utilising the cnot gate in the centre of the chip . however , it is clear from the characterisation of the device that the beam splitter reflectivities deviate from their ideal values and even for perfectly indistinguishable photons we can not expect to be able to prepare precisely these states . we therefore also calculated the theoretically expected modified density matrices for the purpose of the current benchmark test . similarly to the markov chain monte carlo - based calibration , we modelled the theoretically expected density matrix taking the real chip parameters into account as ( see [ appendixc ] ) @xmath64\label{twoqubitrho}\ ] ] within the two - qubit subspace , where @xmath30 is chosen such that @xmath65 . the indistinguishable part exhibiting ideal quantum interference is obtained simply as @xmath66 using @xmath67 where the amplitudes of the two distinguishable possibilities are @xmath68 and @xmath69 where the unitary @xmath70 is now only of the preparation stage ( see [ appendixa ] ) . the distinguishable part can be obtained as the statistical mixture of the two distinguishable possibilities @xmath71 in order to arrive at a valid density matrix the two qubit state has to be normalised to account for the fact that the cnot works probabilistically , i.e. the state is projected to the two - qubit subspace . note that the diagonals of @xmath72 and @xmath73 consist of the familiar looking elements : in the former case probabilities are added , and in the latter case amplitudes are added . to benchmark the photonic chip we reconstructed the density matrix by quantum state tomography @xcite . instead of the more commonly used maximum likelihood tomography we used bayesian mcmc method . this serves two purposes : it allows us to conveniently ensure that the density matrix is physical and to obtain rigorous error bars for the density matrix fidelity against theoretical expectations . to obtain the required experimental data , we used nine different phase settings on the output side phase shifters per input state , which are in principle enough to characterise the state fully @xcite . that is , we measured the qubits along @xmath74 as accurately as possible . each one of the measurements gives information not only about the corresponding two - qubit density matrix element but also about the single - qubit terms . however , to account for imperfections and finite number of repetitions we resorted to numerical methods . we parametrised the density matrix as @xmath75 , where @xmath76 s are the 15 free real unknown parameters , with @xmath77 , and @xmath78 are the pauli matrices including the identity @xmath79 in the notation . similar to parameter estimation , we can obtain the distribution of the density matrix parameters @xmath80 using @xmath81 or @xmath82 up to a constant . here @xmath83 denotes the collective set of observations . we can write the multinomial likelihood as @xmath84 where d is a constant , the index @xmath85 runs over the nine tomography phase settings , @xmath86 is the number of times that we detected the qubit state @xmath87 and @xmath88 is the corresponding expected probability . these probabilities depend on both the unknown density matrix parameters that we optimise over and the currently known phase settings of the shifters 5 - 8 . we used a uniform prior @xmath89 over physical density matrices , i.e. @xmath89 is constant whenever the corresponding eigenvalues of @xmath90 are non - negative and zero otherwise . in practice this means that @xmath91 is set to a large value ( ideally infinite ) whenever the mcmc algorithm attempts to move outside the physical region . there are numerous other alternative ways to choose the prior @xcite . we chose the present one due to numerical convenience . the real part of measured density matrices of all the four bell states are shown in figure [ fig3 ] . the fidelity estimates and the mean density matrices are obtained by averaging over the random walk . both the fidelity against the ideal bell state and the fidelity against the predicted density matrix taking into account the calibrated parameters are shown . these are denoted by @xmath92 and @xmath93 , respectively . we used the definition of fidelity @xcite @xmath94 where @xmath95 is the experimentally measured density matrix . to calculate the fidelity @xmath96 we choose the reference density matrix @xmath97 to be the density matrix of one of the bell states . to obtain @xmath98 we set the reference density matrix to the density matrix calculated using eq . ( [ twoqubitrho ] ) with the same settings for the phase shifters as for the bell states but using real coupler efficiencies and taking the non - ideal photon indistinguishability into account . for the four bell states @xmath92 is [email protected]@xmath1 on average this is an improvement over previously reported @xcite fidelities . the remaining 5 - 6@xmath1 imperfection mostly arises from the non ideal reflectivities and the photon source , as the average fidelity increases to [email protected]@xmath1 for @xmath93 . this indicates that a major source of the decrease of the fidelity are the non - ideal reflectivities of the on - chip directional couplers . generation and measurements of bell states . ( a)-(d ) : real part of density matrix of bell states @xmath99,@xmath100 , @xmath101 and @xmath102 respectively . the corresponding phase shifts were @xmath103 , @xmath104 , @xmath105 and @xmath106 for @xmath99 ; @xmath103 , @xmath104 , @xmath107 and @xmath106 for @xmath100 ; @xmath108 , @xmath104 , @xmath107 and @xmath106 for @xmath101 ; @xmath103 , @xmath104 , @xmath105 and @xmath106 for @xmath102 . the two fidelity values at the top of each figure were calculated against the ideal bell states and best theoretical prediction . the quoted fidelity values indicate the bayesian 95% confidence intervals . the insets illustrate typical mcmc fidelity histograms for the ideal ( solid ) and the best theoretical expectation ( dashed).,scaledwidth=100.0% ] to further demonstrate the agreement with the theoretical model and the experimental results , we varied the relative delay of the two photons and performed quantum state tomography for each fixed delay . this experiment can be viewed as a hong - ou - mandel measurement for bell states . depending on the delay between the two photons , the four two - qubit states were expected to change between maximally entangled and totally mixed as predicted by the model presented above . we compared the measured density matrices against ( i ) the ideal bell states , ( ii ) against the best expected density matrix with finite but optimal @xmath25 with real reflectivities and ( iii ) finally against the delay dependent @xmath109 ( with real reflectivities ) , where @xmath25 was deduced from the independent measurement in fig . [ fig2 ] . figure [ fig4 ] ( a)-(d ) illustrate the results for @xmath99 , @xmath100 , @xmath101 and @xmath102 , respectively . there are three curves for each state : the lowest ( red ) curve corresponds to case ( i ) , the middle ( blue ) corresponds to case ( ii ) and top ( magenta ) corresponds to case ( iii ) . in cases ( i ) and ( ii ) one can clearly see how the fidelities peak in analogy with the hong - ou - mandel effect . comparing the experiments with the detailed model results in an increase in the fidelities . the tops of the four peaks correspond to fig . [ fig3 ] . case ( iii ) shows the agreement of theory and experiment most clearly ; the fidelity of the measured density matrices as a function of delay agrees almost perfectly with the model for the delay dependent @xmath110 which is a statistical mixture of distinguishable and indistinguishable behaviour . how the reconstructed density matrices evolve as a function of the delay is shown in a supplementary video . bell states fidelity as a function of the relative delay of the two photons . red : @xmath92 , fidelity against ideal bell states . blue : @xmath93 , fidelity against modified expectation with fitted reflectivities and probability of distinguishability . magenta : fidelity against the best theoretical expectation using fitted reflectivities and the delay - dependent probability of distinguishability . this delay - dependent probability is obtained from the independent hong ou mandel dip scan in fig . [ fig2].,scaledwidth=100.0% ] in conclusion , we have presented an accurate calibration method for a reconfigurable two - qubit quantum photonic chip based on bayesian updating and markov chain monte carlo simulation . an alternative approach to tomography of linear optics circuits has been presented in @xcite . our method differs in several points . additional to employing a statistical inference method we use an explicit model of the device and do not look for a general unitary . this reduces the number of parameters that need to be estimated . furthermore , our method introduces minimal disturbance from calibration to the actual measurements as it is not necessary to switch the input ports for the photons . having obtained the real chip and source parameters we checked the fidelity of a prepared state against a theoretical prediction using these parameters . a maximum fidelity of [email protected]@xmath1 against the theory prediction using the real device parameters shows that practically all error sources are accounted for and improving the photon source as well as the chip fabrication will make it possible to prepare states with fidelities larger than @xmath111 . a remarkable fact is that the almost unit fidelity also extends to mixed states . this enables us to reliably prepare not only maximally entangled states but also mixed states with a varying degree of mixedness . a source of such states can be of considerable interest in investigating mixed state quantum computation and test concepts like quantum discord .
in an earlier work @xcite we had developed a methodology for the calculation of configuration averaged optical conductivity of a disordered alloy based on the augmented space formalsim developed by one of us @xcite . the method was based on the kubo - greenwood formula and we had shown that disorder scattering renormalizes both the green functions and the current terms . in linear response theory , at zero temperature , the generalized susceptibility of a disordered alloy is given by the kubo formula : @xmath0 where , @xmath1 is the vector potential , and @xmath2\rangle\ ] ] @xmath3 is the current operator and @xmath4 is the heaviside step function . if the underlying lattice has cubic symmetry , @xmath5 = @xmath6 . the fluctuation dissipation theorem then relates the imaginary part of the laplace transform of the generalized susceptibility to the laplace transform of a correlation function : @xmath7 where , @xmath8 since the response function is independent of the direction label @xmath9 for cubic symmetry , in the following we shall drop this symbol . in case of other symmetries we have to generalize our results for different directions . our goal will be , given a quantum hamiltonian " * h * to obtain the correlation function , @xmath10 we shall now determine the correlation directly via the recursion method as described by viswnath and mller@xcite . in order to simplify the expressions for the dynamical quantities as produced by the hamiltonian , we consider henceforth the modified hamiltonian @xmath11 , whose band energy is shifted to zero . let @xmath12 the time evolution of this _ bra _ is goverened by the schrdinger equation @xmath13 we shall now generate an orthogonal basis @xmath14 for representation of equation ( [ eq2 ] ) . we shall do this in the following way : 1 . we begin with initial conditions : @xmath15 2 . we now generate the new basis members by a three term recurrance relationship : @xmath16 + where , @xmath17 we now expand the bra @xmath18 in this orthogonal basis : @xmath19 substituting equation ( [ eq3 ] ) into equation ( [ eq2 ] ) and using orthogonality of the basis , we get : @xmath20 with @xmath21 and @xmath22 . we shall now show that the pair of sequences generated by us , namely , @xmath23 and @xmath24 are enough for us to generate the correlation function . we note first that : @xmath25 let us define the laplace transform : @xmath26 putting this back in equation([eq4 ] ) we get : @xmath27 this set of equations can be solved for @xmath28 as a continued fraction representation : @xmath29 the structure function , which is the laplace transform of the correlation function can then be obtained from the above : @xmath30 augmented space method was first proposed by one of us @xcite as a feasible technique for carrying out configuration averaging in disordered systems . in particular , it has been applied extensively for electronic structure of disordered alloys . the readers are referred to the article @xcite for details . here we shall discuss the salient features of this method for application to configuration averages of correlation functions . the problem addressed involves the configuration averaging of a function of many independent random variables , e.g. @xmath31 where @xmath32 labels a set of lattice points . the first step is to associate with each random variable @xmath33 an operator @xmath34 such that the spectral density of this operator is the probability density of the random variable : @xmath35 for example , if @xmath33 takes the values 0 and 1 with probabilities @xmath36 and @xmath37 then the operator * m@xmath38 is of rank 2 and acts on the configuration space of @xmath33 , @xmath39 , spanned by the eigenstates @xmath40 and @xmath41 of m@xmath38 . the state @xmath42 and the representation of m@xmath38 in the basis @xmath43 is @xmath44 the augmented space theorem then states that : @xmath45 where @xmath46 is an operator of rank @xmath47 obtained by replacing each @xmath33 by m@xmath38 . this operator acts on the configuration space of the set of random variables , @xmath48 = @xmath49 . a state in this space is given by a pattern of configurations at each site , i.e. sequences like @xmath50 which can be uniquely represented by the sequence of sites where we have ( say ) a @xmath51 configuration . this sequence is called a _ cardinality sequence_. the empty cardinality sequence , @xmath52 , is the pattern of @xmath53-s everywhere . if we go back to equations ( 2)-(5 ) , we notice that for a disordered binary alloy , @xmath54 $ ] . the augmented space theorem then states that : @xmath55\ ] ] where the augmented space hamiltonian and the current operators are constructed by replacing every random variable @xmath33 by the corresponding operator m@xmath38 . the recursion may now be modified step by step in the full augmented space : @xmath56 the time evolution of this _ bra _ is goverened by the schrdinger equation @xmath57 as before , we shall generate the orthogonal basis @xmath14 for representation of equation ( [ eq2a ] ) : 1 . we begin with initial conditions : @xmath58 2 . the new basis members are generated by a three term recurrance relationship : @xmath59 + where , @xmath60 we now expand the bra @xmath18 in this orthogonal basis : @xmath61 continuing exactly as in the last section we get , @xmath62 taking laplace transforms and using the three term recurrance , @xmath63 the configuration averaged structure function , which is the laplace transform of the averaged correlation function can then be obtained from the above : @xmath64 the equations ( 11)-(14 ) will form the basis of our calculation of the configuration averaged correlation function . let us begin with a hamiltonian for a random binary alloy represented in the basis of a tight - binding linear muffin - tin orbitals method ( tb - lmto ) @xmath65 @xcite . for the sake of notational clarity we have suppressed the angular momentum index @xmath66 . in what follows , we may consider @xmath32 to be a composite index @xmath67 . @xmath68 @xmath69 and t@xmath70 are projection and transfer operators on the space spanned by the tb - lmto basis . @xmath71 and @xmath72 are the tb - lmto potential parameters and @xmath73 the structure matrix . for a binary alloy the former are random and may be given by @xmath74 here the random variables @xmath33 takes values 0 and 1 with probabilities @xmath36 and @xmath75 , as before . the augmented space theorem then builds up the augmented hamiltonian by replacing the binary random variables @xmath76 by the corresponding operators @xmath77 described in the earlier section . @xmath78 where the configuration space operators are : @xmath79 the augmented space hamiltonian is then given by : @xmath80 next we look at the expression for the current operator in the tb - lmto basis : @xmath81 the dot refers to derivative with respect to energy . in this basis , the matrix elements of the current operator can be written as @xmath82\nonumber\\\end{aligned}\ ] ] where @xmath83 the technique for calculating these matrix elements has been described in detailed by hobbs @xcite . we have also used this technique in our earlier paper @xcite and we shall use it here as well . the readers are referred to these two papers for details . ideally the next step would be to calculate @xmath84 as the current terms between two sites when they are occupied by atom pairs aa , bb , ab and ba embedded in the disordered medium . a simpler first step would be to obtain these current terms from the pure a and b and from the ordered ab alloy . in general , the current operator can be written as : ^= _ r j^(0 ) p_r + _ r_r j^ ( ) t_rr where @xmath85 . in a disordered alloy , the current representations are random : @xmath86 using the augmented space theorem , we replace the random variable @xmath33 by an operator @xmath87 in the expression of @xmath88 we get the current operator in the augmented space : @xmath89\otimes { \mathbf t}_{r , r'}\right ) + xy\ j^\mu_3(\chi)\ { \mathbf t}_{r}^{\uparrow\downarrow}\otimes{\mathbf t}_{r'}^{\uparrow\downarrow}\otimes { \mathbf t}_{r , r'}\nonumber\\\end{aligned}\ ] ] where @xmath90\\ j^\mu_2(0)=\sqrt{xy}[j_{aa}^\mu(0)-j_{bb}^\mu(0)]\\ \langle j(\chi)\rangle = x^2j_{aa}^\mu(\chi)+xy\ \left(j_{ab}^\mu(\chi)+j_{ba}^\mu(\chi)\right)+y^2j_{bb}^\mu(\chi)\\ j_1^\mu(\chi)=x\left(j_{aa}^\mu(\chi)-j_{ab}^\mu(\chi)\right)+y\left(j_{ba}^\mu(\chi)-j_{bb}^\mu(\chi)\right)\\ j_2^\mu(\chi)=x\left(j_{aa}^\mu(\chi)-j_{ba}^\mu(\chi)\right ) + y\left(j_{ab}^\mu(\chi)-j_{bb}^\mu(\chi)\right)\\ j_3^\mu(\chi)=j_{aa}^\mu - j_{ab}^\mu - j_{ba}^\mu+j_{bb}^\mu\end{aligned}\ ] ] this augmented current operator is used to construct the starting state of the recursion as described in the last section . for application of our methodology we have chosen two alloy systems disordered : cuau ( 50 - 50 ) and disordered agau ( 50 - 50 ) . our choice was governed by two considerations : first , we had studied these alloys in an earlier work with a different approach @xcite and it would be interesting to compare the two approaches ; and secondly , the fact that we have experimental data available for both these alloys and it would be interesting to compare our theoretical estimates with experiment . we have begun our study with the self - consistent tb - lmto - asr calculation on agau and cuau 50 - 50 alloys . we have minimized the energy with respect to the variation in the average lattice constant for both the alloys . we have calculated the current terms from the basis functions and potential parameters obtained from a tb - lmto calculation for the ordered 50 - 50 alloys . ideally we should have embedded aa , bb and ab pairs in the disordered medium and obtained the current expressions from such a calculation . however , in our earlier work @xcite we have shown that effect of randomness in the current terms is small . this gives us confidence to proceed with these estimates . to check the density of states for the disordered alloys we have used our tb - lmto - asr programme developed by us @xcite . the density of states has been shown in our earlier work @xcite . the main features of interest is that the filled @xmath91-states lie about 2 ev below the fermi level for both these alloys . = 2.5 in = 2.5 in + = 2.5 in = 2.5 in + in figure [ fig2 ] we show the continued fraction coefficients for a few different angular momentum components . the coefficients do show a tendency of showing oscillatory convergence , but within the maximum ten steps that we were able to go , it was not possible to accurately locate the converged values . we therefore chose not to use the square - root terminator , but opted for the beer - pettifor terminator @xcite . = 3.5 in experimental data on disordered cuau ( 50 - 50 ) are available @xcite . the authors have reported high resolution optical data for both ordered and disordered cuau for photon energies between 1.2 ev and 6 ev . in the figure [ fig3 ] , we also show the experimental data for @xmath92 in dotted lines and compare it with the theoretical data in figure [ fig3](bold lines ) . drude like behaviour is clearly seen till photon energies of @xmath93 2.1 ev in theory and around 2 ev in the experiment . subsequently @xmath92 rises again because of transitions from the @xmath91-states . both the theory and experiment show a decrease around 6 ev . the experimental curve shows a sharper dip and the subsequent peak is at a lower energy . we have experimental data on agau ( 50 - 50 ) @xcite , whose density of states closely resembles cuau . the inter - band contribution to the imaginary part of the dielectric function @xmath94 may be obtained from the optical conductivity data , by subtracting away the drude contribution and dividing the result by @xmath95 : @xmath96 . below the onset of the inter - band transitions , this quantity vanishes . it shows a shoulder around 2.5 ev and then reaches a maximum at around 3.5 ev before decreasing and showing a feature around 5.5 ev . = 3.5 in as in the coherent potential approximation ( cpa ) estimates of banhart @xcite there is a @xmath97 0.5ev shift outwards in the theoretical estimate . in the above figure we have shifted our theoretical data by 0.5ev downwards for comparison . banhart attributes this shift to the foundations of the electronic theory used : local density approximation which can not reproduce excited states properly . this had been discussed also by fehrenbach and bross @xcite . to this we shall add that it is doubtful whether the energy linearized tb - lmto can accurately reproduce the large energy window of our calculation . it may be useful in future to base our calculations on the newly developed nmto formalism @xcite . in addition , the termination procedure of the recursion requires further refinement , as the convergence of the continued fraction coefficients is not as rapid as in density of states calculations . banhart notes that the cpa neglects all short - ranged order . however , since our calculations in the augmented space recursion does include correlated scattering from different sites , we can not attribute the discrepancy to the cpa . apart from the energy shift , the experimental data are in reasonable agreement with our theory . the general shape with a maximum is clearly reproduced . the feature shown in the experimental data around 5.5 ev is also seen in our theoretical estimate , slightly shifted in energy . the shoulder in the experimental data just above 2 ev is not properly reproduced in the theoretical curve . the cpa estimate of banhart ( @xcite ) also does not seem to show such a prominent shoulder . is it possible that the experimental alloy had short ranged order ? it is known that short - ranged order can have influence on optical properties ( @xcite ) . the augmented state formalism can handle short - ranged order ( @xcite ) and we shall study this effect in a future communication . one of us ( kt ) would like to acknowledge financial support from the csir , india . 99 saha k.k . and mookerjee a. , ( 2005 ) 4559 mookerjee a. , ( 1973 ) l205 ; 1340 viswanath v.s . and mller g. , ( 1993 ) in _ the user friendly recursion method _ " , ( troisieme cycle de la physique , en suisse romande ) mookerjee a. , ( 2003 ) in _ electronic structure of alloys , surfaces and clusters _ " ed . a. mookerjee and d.d . sarma , taylor francis , uk . andersen o.k . and jepsen o. , ( 1984 ) 2571 hobbs d. , piparo e. , girlanda r. and monaca m. , ( 1995 ) 2541 chakrabarty a. and mookerjee a. , ( 2005 ) e. physical j b 44 * 21 beer n. and pettifor d.g . , in _ the electronic structure ofcomplex systems _ " ed . p. phariseau and w.m . temmerman , nato , asi series b * 113 * 769 scott w.r . and muldawer l.,(1976 ) 4426 banhart j. , ( 1999 ) 2139 nielsson p.o . , ( 1970 ) _ phys . * 11 * 1 fehrenbach g.m . and bross h. , ( 1966 ) _ phys . stat . solidi _ ( b ) * 193 * 231 andersen o.k . and saha - dasgupta t. , ( 2000 ) r16219 andrews j.b . , natasi - andrews r.j . and hummel r.e . , ( 1980 ) 1837 mookerjee a. and prasad r. , ( 1993 ) 17724 ; saha t. , dasgupta i. and mookerjee a. , ( 1994 ) 13267	we present here a calculation of the configuration averaged optical conductivity of random binary alloys cuau and agau . our formulation is based on the augmented space formalism proposed by mookerjee [ j. phys . c : solid state phys . * 6 * 1340 ( 1973 ) ] and the optical conductivity is obtained directly through a recursive procedure suggested by viswanath and mller [ the user friendly recursion method " , troisieme cycle de la physique , en suisse romande ( 1993 ) ] . # 1 # 1 # 1e - mail : # 1 # 1to appear in # 1 # 1submitted to # 1
bipartite and multipartite entanglement is considered a basic resource in most applications of quantum information , communication , and technology ( see for instance refs . entanglement is fragile , and it is well - known that in some cases interactions with an environment external to the entangled systems may decrease the quantum correlations , degrading this valuable resource @xcite . momentum acts as a very special environment which every particle possesses and can not get rid of . consider for instance a bipartite system which initially is spin - entangled and with the momentum distributions factorized . it will decrease its spin - spin correlations provided any of the two particles entangles its spin with its momentum . this simple idea was studied in the natural framework of special relativity , where changing the reference frame induces wigner rotations that entangle each spin with its momentum @xcite . however , this is just a kinematical , frame - dependent effect only , and not a real dynamical interaction . on the other hand , this type of reasoning is also related to which - path detection @xcite . in addition , an experiment observing photon polarization disentanglement by correlation transfer , in this case to the photon s position , was performed @xcite . can local interactions entangling spin with momentum produce the loss of non - local spin - spin entanglement ? to our knowledge this question has not been explored in the literature . more remarkably , any particle owns a certain momentum distribution acting as an intrinsic environment , which can never be eliminated by improving the experimental conditions . but how does this fact affect the spin - spin correlations ? this is the question we want to analyze in this paper . in sec . [ lbsetc2 ] we consider a bipartite system , composed by two @xmath0 fermions or two photons , which are initially in a bell spin state @xmath1 . we use a formalism @xcite that shows the decrease of spin entanglement whenever an interaction locally entangling spin with momentum takes place . we obtain the negativity @xmath2 @xcite in terms of an integral depending on the spin rotation angle conditional to the momentum . in sec . [ lbsetc3 ] we analyze this physical phenomenon in two specific situations : ( i ) two spin-@xmath3 fermions in a @xmath1 bell state , with gaussian momentum distributions , that fly apart while one of them passes through a local magnetic field . their spin entanglement decreases as a consequence of the transfer of correlations to the momentum of the latter fermion . and ( ii ) two photons in a polarization @xmath1 bell state , with gaussian momentum distributions , which separate while one of them traverses an optically - active medium . this medium will entangle the polarization with the momentum and thus decrease the polarization entanglement . this is an unavoidable source of decoherence . in sec . [ lbsetc4 ] we show that the apparent purely quantum communication resulting from these procedures is not possible . classical communication has to be exchanged for the protocol to operate . the paper ends with our conclusions in sec . [ lbsetc5 ] . we consider a maximally spin - entangled state for two @xmath0 fermions @xmath4 and @xmath5 , or two photons @xmath4 and @xmath5 . the case we analyze is that in which the two particles are far apart already . this avoids dealing with symmetrization issues . indeed , our state is the maximally entangled one for two @xmath0 spins or polarizations , containing 1 ebit . @xmath6 where @xmath7 and @xmath8 are the corresponding momentum vectors of particles @xmath4 and @xmath5 , and @xmath9 with bimodal momentum distribution @xmath10 . @xmath11 and @xmath12 are the two momentum values considered associated to particle @xmath4 , @xmath13 , @xmath14 , on the one hand , and particle @xmath5 , @xmath15 , @xmath16 , on the other hand . we consider for the time being this kind of distribution for illustrative purposes . at the end of this section we generalize our results to arbitrary momentum distributions of the two particles . @xmath17 and @xmath18 represent either spin vectors pointing up and down along the @xmath19-axis , in the fermionic case , or right - handed and left - handed circular polarizations , in the photonic case . if we trace out the momentum degrees of freedom in eq . ( [ deco1 ] ) , we obtain the usual spin bell state , @xmath20 . we consider a local interaction which entangles the spin of each particle with its momentum through a real unitary ( orthogonal ) transformation @xmath21 . we choose a real transformation for the sake of simplicity and in order to obtain fully analytical results . the generalization for inclusion of complex phases is straightforward but adds nothing of relevance in this section . we will take it fully into account in sec . [ lbsetc3 ] . each state vector in eq . ( [ deco2 ] ) transforms as @xmath22=\left ( \begin{array}{cccc } \cos \theta_{\vec{p_1 } } \\ \sin \theta_{\vec{p_1 } } \\ \end{array } \right ) \frac{\delta_{\vec{p}\vec{p_1}}}{\sqrt{2}}+\left(\begin{array}{cccc } \cos \theta_{\vec{p_2 } } \\ \sin \theta_{\vec{p_2 } } \\ \end{array } \right)\frac{\delta_{\vec{p}\vec{p_2}}}{\sqrt{2 } } , \nonumber \\ & & \psi_\downarrow(\vec{p})=\left ( \begin{array}{cccc } 0 \\ { \cal m } ( \vec{p } ) \\ \end{array } \right ) \rightarrow\nonumber\\ & & u[\psi_\downarrow(\vec{p})]=\left ( \begin{array}{cccc } -\sin \theta_{\vec{p_1 } } \\ \,\,\,\,\ , \cos \theta_{\vec{p_1 } } \\ \end{array } \right ) \frac{\delta_{\vec{p}\vec{p_1}}}{\sqrt{2}}+\left(\begin{array}{cccc } -\sin \theta_{\vec{p_2 } } \\ \,\,\,\,\ , \cos \theta_{\vec{p_2 } } \\ \end{array } \right ) \frac{\delta_{\vec{p}\vec{p_2}}}{\sqrt{2 } } , \nonumber\\ \label{deco3}\end{aligned}\ ] ] where @xmath23 and @xmath24 produce a spin - momentum entangled state whenever @xmath25 . the effect of this local interaction is that a part of the non - local spin - spin entanglement is transferred to the spin - momentum one , and the degree of entanglement of the spins decreases . to show this , we consider the state ( [ deco1 ] ) evolved with the transformation @xmath21 , and trace out the momenta . @xmath26\{u^{(\rm a)}[\psi^{(\rm a)}_{s ' } ( \vec{p_a})]\}^{\dag } ) \nonumber\\&&\otimes \mathrm{tr}_{\rm \vec{p_b}}(u^{(\rm b)}[\psi^{(\rm b)}_{-s } ( \vec{p_b } ) ] \{u^{(\rm b)}[\psi^{(\rm b)}_{-s ' } ( \vec{p_b})]\}^{\dag}),\label{deco4}\end{aligned}\ ] ] where @xmath27 . it can be appreciated in eq . ( [ deco4 ] ) that the expression is decomposable in sum of tensor products of 2@xmath282 spin blocks , each corresponding to each particle . we compute now the different blocks , corresponding to the four possible tensor products of the states ( [ deco3 ] ) @xmath29 & = & \frac{1}{2 } \left(\begin{array}{cc}c_1 ^ 2+c_2 ^ 2 & c_1 s_1+c_2 s_2\\c_1 s_1+c_2 s_2 & s_1 ^ 2+s_2 ^ 2\end{array}\right),\nonumber \\ \mathrm{tr}_{\rm \vec{p}}[u\psi_\uparrow(u\psi_\downarrow)^{\dag } ] & = & \frac{1}{2}\left(\begin{array}{cc}-c_1 s_1-c_2 s_2 & c_1 ^ 2+c_2 ^ 2\\-s_1 ^ 2-s_2 ^ 2 & c_1 s_1+c_2 s_2\end{array}\right),\nonumber\\ \mathrm{tr}_{\rm \vec{p}}[u\psi_\downarrow(u\psi_\uparrow)^{\dag}]&= & \frac{1}{2}\left(\begin{array}{cc}-c_1 s_1-c_2 s_2 & -s_1 ^ 2-s_2 ^ 2\\c_1 ^ 2+c_2 ^ 2 & c_1 s_1+c_2 s_2\end{array}\right),\nonumber\\ \mathrm{tr}_{\rm \vec{p}}[u\psi_\downarrow(u\psi_\downarrow)^{\dag}]&=&\frac{1}{2 } \left(\begin{array}{cc}s_1 ^ 2+s_2 ^ 2 & -c_1 s_1-c_2 s_2\\-c_1 s_1-c_2 s_2 & c_1 ^ 2+c_2 ^ 2\end{array}\right),\nonumber\end{aligned}\ ] ] where @xmath30 and @xmath31 . this way , it is possible to compute the effects of the local interaction @xmath21 in the state @xmath32 after tracing out the momentum . we choose equal interaction angles for the two particles , @xmath33 , as a natural simplification . the resulting bipartite spin state is @xmath34 where @xmath35 and @xmath36 . the entanglement monotone we will use is the negativity @xcite , defined as @xmath37 , where @xmath38 is the smallest eigenvalue of the partial transpose ( pt ) matrix of eq . ( [ deco6 ] ) . it is very easily computable , and is found to be @xmath39 from this expression it can be appreciated that for @xmath40 the entanglement remains maximal ( 1 ebit ) , and for @xmath23 separating from @xmath24 the entanglement decreases , until @xmath41 , where it vanishes and the state becomes separable . we plot this behavior in figure [ grafdeco ] , showing the negativity @xmath2 in eq . ( [ deco7 ] ) as a function of @xmath23 and @xmath24 . every local interaction producing spin - momentum entanglement will in general diminish the initial maximal spin - spin entanglement of the two particles , thus degrading this resource . this result is valid either for @xmath0 fermions or photons , as they both have a two - dimensional internal hilbert space . in eq . ( [ deco7 ] ) as a function of @xmath23 and @xmath24.[grafdeco],width=264 ] the generalization of eq . ( [ deco7 ] ) to a uniform distribution with @xmath42 different momenta is straightforward , and the spectrum of the corresponding pt matrix is @xmath43\right\},\label{deco8}\ ] ] with a resulting negativity @xmath44 we take now the continuous limit , for an arbitrary momentum distribution @xmath45 for each particle . we suppose for the sake of simplicity @xmath46 , although the spatial distributions do not overlap , as the two particles are far away . accordingly , @xmath2 will be @xmath47 notice that this expression involves an integration over the momentum variables @xmath48 and @xmath49 , associated to the same particle ( not to each of them ) . we point out that , according to ( [ deco10 ] ) , in the case where momentum does not entangle with spin ( i.e. whenever @xmath50 is a constant ) , then @xmath51 and thus the spin - spin entanglement remains maximal . otherwise , the spin - spin entanglement would decrease due to the transfer of correlations to the spin - momentum part . the loss of spin entanglement under a spin - momentum entangling transformation can take place in a variety of possible situations . wigner rotations that appear under relativistic change of reference frame entangle each spin with its momentum producing loss of spin - spin entanglement @xcite . this is just a kinematical - relativistic effect , not due to a dynamical interaction . in the rest of the paper we focus on two relevant examples of these interactions , taking fully into account the complex phases : a local homogeneous magnetic field , for the two - fermion case , and a local optically - active medium , for the two - photon case in this section we analyze a bipartite system , composed by two @xmath0 neutral fermions @xmath4 and @xmath5 , which are initially far apart and in a bell spin state @xmath1 , with factorized gaussian momentum distributions . we consider that one of them traverses a region where a finite , homogeneous magnetic field exists . as a result , it will transfer part of its spin correlations to the momentum . the initial spin - entangled state for the two fermions @xmath4 and @xmath5 is @xmath52 where @xmath7 and @xmath8 are the corresponding momentum vectors of particles @xmath4 and @xmath5 , and @xmath53 with gaussian momentum distribution @xmath54 $ ] . in eqs . ( [ v2deco2 ] ) we are not indicating explicitly the particle index . in the center of mass frame , @xmath55 , and we consider that the two particles are flying apart from each other . @xmath17 and @xmath18 represent spin vectors pointing up and down along the @xmath19-axis , respectively . if we trace out momentum degrees of freedom in eq . ( [ v2deco1 ] ) , we obtain the usual spin bell state , @xmath20 . suppose a local interaction which entangles the spin of fermion @xmath4 with its momentum through a unitary transformation @xmath21 . in this case we choose a magnetic field @xmath56 which is constant on a bounded region @xmath57 of length @xmath58 , along the direction of @xmath59 , vanishes outside @xmath57 , and extends infinitely with a constant value along the other two orthogonal directions , as shown in figure [ v2grafdeco1 ] . we take @xmath56 along the direction orthogonal to @xmath59 , so it is divergenceless , @xmath60 , and we quantize the spin along @xmath56 so that @xmath61 , with @xmath62 the corresponding spin component . due to the rotational invariance of the spin singlet , this choice is completely general . in momentum space , the problem reduces to one dimension , the one associated to the direction of @xmath59 . we will denote from now on @xmath63 to the corresponding momentum coordinate . traverses a constant magnetic field @xmath56 located in region @xmath57 with a width @xmath58 along the direction of @xmath59 . [ v2grafdeco1],width=207 ] the system hamiltonian can be written as @xmath64 where @xmath65 is the magnetic moment of our neutral particle . accordingly , we get @xmath66&=&\left(-\gamma\ , ( \vec{b}_0\cdot \vec{s})\ , [ \delta(x)-\delta(l - x)],0,0\right),\label{v3deco2}\\ \dot{\vec{s}}\,=\ , i[h,\vec{s}]&= & -\gamma\,(\vec{b_0}\wedge\vec{s})\ , \theta(x)\,\theta(l - x),\nonumber\\ \ddot{\vec{s}}\,=\ , i[h,\dot{\vec{s}}]&= & -\gamma\,\left[(\vec{b_0}\cdot\vec{s})\,\vec{b_0}- \vec{b_0}^2 \vec{s}\right]\theta(x)\,\theta(l - x)\nonumber\\&&-\,i \gamma\,(\vec{b_0}\wedge\vec{s } ) \left(\frac{\vec{p}}{m } [ \vec{p},(\delta(x)-\delta(l - x))]\right).\nonumber\end{aligned}\ ] ] from the first of the above equations we obtain @xmath67 $ ] like using matching conditions at @xmath68 . the second and third equations give the spin evolution . by inspection we see that i ) the spin remains parallel to @xmath69 if it was initially so and , ii ) the spin is constant in this case @xmath70 . hence , in spite of choosing a case where the spin is conserved , its entanglement with the momentum decreases the spin correlations with the idle fermion . the effect of the magnetic field on particle @xmath4 can be seen in its state . behind the region @xmath57 , the resulting state vector as transformed from the one in eq . ( [ v2deco2 ] ) is @xmath71={\cal{t}}_{\uparrow}(p)\left ( \begin{array}{cccc } { \cal g } ( \vec{p } ) \\ 0 \\ \end{array } \right ) , \nonumber \\ & & \psi_\downarrow(\vec{p})\rightarrow u[\psi_\downarrow(\vec{p})]={\cal{t}}_{\downarrow}(p)\left ( \begin{array}{cccc } 0 \\ { \cal g } ( \vec{p } ) \\ \end{array}\right ) , \nonumber\\ \label{v2deco3}\end{aligned}\ ] ] where @xmath72 ( @xmath73 ) is the transmission coefficient associated to the mesa ( well ) potential induced by @xmath56 , for initial spin up ( down ) . it is given by @xmath74 where @xmath75 , as given by ( [ v3deco2 ] ) , @xmath76 and @xmath77 for spin @xmath78 . as expected for @xmath79 , @xmath80 . the initial state is preserved so no spin - momentum correlations are generated . in general , for @xmath81 , @xmath82 , producing spin - momentum entanglement . we are considering here just transmission through the region @xmath57 , without taking into account the reflection of the wave packets . we suppose all the measurements will take place beyond @xmath57 so we may normalize the final transmitted state to 1 . finally , the net effect of this local interaction is the reshuffling of spin - momentum correlations in the state of the active fermion . accordingly , the degree of spin - spin entanglement between both particles decreases . as was done in eq . ( [ deco4 ] ) , we evolve the state ( [ v2deco1 ] ) with the transformation @xmath21 and trace out the momenta @xmath83\{u[\psi^{(\rm a)}_{s ' } ( \vec{p_a})]\}^{\dag } ) \nonumber\\&&\otimes \mathrm{tr}_{\rm \vec{p_b}}(\psi^{(\rm b)}_{-s } ( \vec{p_b } ) \{\psi^{(\rm b)}_{-s ' } ( \vec{p_b})\}^{\dag}),\label{v2deco4}\end{aligned}\ ] ] where @xmath27 . the traces corresponding to particle @xmath5 give just the initial spin states , @xmath84 , @xmath85 , @xmath86 , @xmath87 , because @xmath21 is just the identity for @xmath5 . the resulting , properly normalized spin - spin state is @xmath88 where @xmath89 the negativity for this state is found to be @xmath90 this expression for @xmath2 is rather illuminating and its behavior easy to understand . for the initial state ( [ v2deco1 ] ) @xmath51 ( 1 initial ebit ) , and , as long as the magnetic field is increased , @xmath72 and @xmath73 become more different , making the term @xmath91 smaller , and diminishing @xmath2 . in eq . ( [ v2deco8 ] ) as a function of @xmath92 for @xmath93 , @xmath94 , @xmath95 , and @xmath96 , 2 , and 3 . the higher curves correspond to the thinner @xmath97 s . all quantities are given in @xmath98 units . [ v2grafdeco2],width=264 ] in eq . ( [ v2deco8 ] ) as a function of @xmath58 for @xmath93 , @xmath94 , @xmath99 , and @xmath100 . all quantities are given in @xmath101 units . [ v2grafdeco3],width=264 ] on the other hand , the wider @xmath102 the more destructive interference between @xmath72 and @xmath73 will occur , reducing @xmath2 . we plot in figure [ v2grafdeco2 ] the negativity @xmath2 as a function of @xmath103 , with @xmath104 , for @xmath93 , @xmath105 , @xmath95 , and @xmath96 , 2 , and 3 . all quantities are given in @xmath98 units . the entanglement goes to zero with increasing @xmath106 , and the wider @xmath102 , the lesser the entanglement . a similar behaviour arises from the cumulative effect of the barrier ; the larger @xmath58 , the smaller the entanglement . we show in figure [ v2grafdeco3 ] this behavior , plotting @xmath2 as a function of @xmath58 for @xmath93 , @xmath94 , @xmath99 , and @xmath100 . in this section we analyze a bipartite system , composed by two photons @xmath4 and @xmath5 , which are far apart in a polarization bell state @xmath1 with factorized gaussian momentum distributions . we consider that the photon @xmath4 traverses a region where a finite , optically - active medium , exists . as a result , part of its spin correlations will be transferred to the momentum . basically the mathematical formalism used for the two - fermion case is also valid here , with @xmath107 and @xmath108 indicating right - hand and left - hand circular polarizations , which we will denote by r and l indices . the transmission coefficient in the wkb approximation , at lowest order , takes now the form of a complex phase , depending on the polarization @xmath109,\;\;\;s = r , l,\label{v2deco8ph}\ ] ] where the refraction indices are @xmath110 and @xmath111 , @xmath112 are two of the matrix elements of the susceptibility @xmath113 , @xmath114 @xmath113 is produced , for example , by an isotropic dielectric placed in a magnetic field @xmath56 directed along @xmath19 , which is also the direction of photon propagation . @xmath58 is the dielectric length that the photon traverses . @xmath111 and @xmath112 are @xmath115,\nonumber\\ & & \chi_{12}(w):=\frac{{\cal{n } } e^2}{m\epsilon_0}\left[\frac{ww_c}{(w_0 ^ 2-w^2)^2+w^2w_c^2}\right],\end{aligned}\ ] ] where the cyclotron frequency @xmath116 , @xmath117 is the electron charge , @xmath118 its mass , @xmath119 the resonance frequency of the optically - active medium , @xmath120 the number of electrons per unit volume , and @xmath121 the vacuum electric permittivity . the next order correction would include factors @xmath122 in the denominators of the transmission coefficients . however , the approximation that considers these factors coincides exactly with the lowest order one when taking into account just linear terms in @xmath56 . we will consider the realistic case in which @xmath123 is small as compared to the photon average energy . thus we will work from the very beginning just with the transmission coefficients ( [ v2deco8ph ] ) . the negativity @xmath2 , obtained for this case in analogy with the two - fermion case , is @xmath124 where @xmath125 is the average momentum of photon @xmath4 , @xmath126 its momentum width , and @xmath127 . we plot in figure [ v2grafdecophoton1 ] the negativity @xmath2 in eq . ( [ v2deco11ph ] ) as a function of @xmath128 for @xmath129 , @xmath130 , and @xmath131 . all quantities are given in @xmath132 units . the entanglement decreases as the magnetic field @xmath133 or the length @xmath58 of the dielectric increase . we plot also in figure [ v2grafdecophoton2 ] the negativity @xmath2 as a function of @xmath97 for @xmath129 , @xmath134 , and @xmath131 . surprisingly , and opposite to the two - fermion case , the entanglement increases as the momentum width @xmath97 of the photon is larger . this effect comes from the fact that for larger widths , centered at @xmath119 , the contribution from the region around the resonance frequency @xmath119 , in which the effect of the medium is appreciable , becomes smaller . on the other hand , in the limit of negligible width , the spin could not get entangled with the momentum so in this limit the spin - spin entanglement would remain maximal . we observe then that there is a region of intermediate widths @xmath97 in which the spin - spin entanglement becomes minimal . finally , we plot in figure [ v2grafdecophoton3 ] the negativity @xmath2 as a function of @xmath119 for @xmath129 , @xmath135 , and @xmath136 . the higher curves correspond to the wider @xmath97 s . these graphics show that the entanglement decreases mainly for resonance frequencies @xmath119 around the average momentum @xmath125 . it also shows the surprising behavior mentioned above : for wider @xmath97 , the entanglement is larger , and the interval of @xmath119 for which the entanglement decreases is wider , as expected according to the previous analysis . in eq . ( [ v2deco11ph ] ) as a function of @xmath128 for @xmath129 , @xmath130 , and @xmath131 . all quantities are given in @xmath132 units . [ v2grafdecophoton1],width=264 ] in eq . ( [ v2deco11ph ] ) as a function of @xmath97 for @xmath129 , @xmath134 , and @xmath131 . all quantities are given in @xmath132 units . [ v2grafdecophoton2],width=264 ] in eq . ( [ v2deco11ph ] ) as a function of @xmath119 for @xmath129 , @xmath135 , and @xmath136 . the higher curves correspond to the wider @xmath97 s . all quantities are given in @xmath132 units.[v2grafdecophoton3],width=264 ] a cautious reader may immediately object that , in principle , our preceding analysis seems to suggest the feasibility of communication through a purely quantum channel , i.e. without classical communication , contrary to all known quantum - informational protocols . let us illustrate this point with the following proposal based in the previous two - fermion case . as usual , alice and bob will be the corresponding observers of each fermion . the joint spin - spin state is given by equation , which immediately drives one to the subsequent reduced spin state for fermion @xmath5 @xmath137 but the quantities @xmath138 depend on the magnetic field @xmath139 ( cf . equations and ) , _ which can be controlled by alice_. this allows them to agree on the following procedure . they agree on communicating with a binary alphabet with classical bits @xmath140 and @xmath141 . if alice were to communicate @xmath140 , she would adjust @xmath139 so that the reduced spin state for bob is , for example , @xmath142 they prepare a statistically significative amount of pairs of fermions under such conditions . then bob , when measuring the spin upon his fermion , will typically obtain spin up in the @xmath143 of the measurements and spin down in the remaining @xmath144 . he thus deduces that alice is sending the bit @xmath140 . on the contrary , if alice were to communicate the bit @xmath141 , she would adjust @xmath139 so that the reduced spin state for bob is , for example , @xmath145 they also prepare a statistically significative amount of pairs and bob performs his measurements . he will detect @xmath143 of them in the spin down state and @xmath144 in the spin up state . he thus deduces that alice is sending the bit @xmath141 . notice that this information transmission is carried out without the assistance of classical communication . the flaw stems from the disregarding of the wave packet reflection in alice s site . this can be seen in two complementary ways . on the one hand , to perform a genuine information transmission in which bob s fermion is actually carrying information encoded by alice , he must be able to discern between those fermions whose pairs have been reflected in alice s barrier potential , so that he can securely discard them ( they are not carrying information at all ) . and this is only possible if alice _ classically _ communicates this information to bob . on the other hand , a detailed calculation taking into account the reflection coefficients , hence without the normalization appearing in the denominator of , shows that bob s reduced spin state will be given by @xmath146 where @xmath147 denotes the corresponding reflection coefficient for spin @xmath62 and momentum @xmath63 . the calculation reveals that bob gains no information whatsoever from alice s decisions , unless she classically informs bob about them . mathematically this can be expressed through the unitary character of the process . if no classical information is exchanged , the evolution is locally unitary ( @xmath148\otimes u^{(b)}[\psi^{(b)}]$ ] ) and thus can not change the entanglement shared by both parties ( the entanglement is invariant under local unitary evolution ) . consequently no information through the purely quantum channel can be obtained . on the contrary , if alice classically sends information to bob , she is actually selecting a subset of her incoming fermions , i.e. she is _ projecting _ her state ( @xmath149 $ ] , where @xmath150 is an orthogonal projector ) , which is a non - unitary operator which changes the entanglement class . this fact allows them to exploit the initial quantum correlations between their fermions to establish a communication protocol . in summary , this example shows once more the impossibility of using quantum correlations , i.e. entanglement to exchange information without the aid of classical communication . we showed the spin entanglement loss by transfer of correlations to the momentum of one of the particles , through a local spin - momentum entangling interaction . this phenomenon , already analyzed for a non - interacting particular case in the context of wigner rotations of special relativity , may produce decoherence of bell spin states . the momentum of each particle is a very simple reservoir and indeed it is one that can not be eliminated by improving the experimental conditions , due to heisenberg s principle . we show that an @xmath0 fermion ( photon ) , which initially belongs to a bell spin state , may lose its spin correlations due to this physical phenomenon when traversing a local magnetic field ( optically - active medium ) . these specific media entangle each component of the spin state of the particle with its momentum , like in a stern - gerlach device . this could have implications for quantum communication and information processing devices . this work was partially supported by the spanish mec projects no . fis2005 - 05304 ( l.l . and j.l . ) and no . fis2004 - 01576 ( d.s . ) . l.l . acknowledges support from the fpu grant no . ap2003 - 0014 .	we show the decrease of spin - spin entanglement between two @xmath0 fermions or two photons due to local transfer of correlations from the spin to the momentum degree of freedom of one of the two particles . we explicitly show how this phenomenon operates in the case where one of the two fermions ( photons ) passes through a local homogeneous magnetic field ( optically - active medium ) , losing its spin correlations with the other particle . # 1#1 # 1 # 1#1\| # 1\|#1 # 1#2#1\|#2 # 1#2#1,#2 # 1#1 s
in microelectromechanical systems ( mems ) , the motion of contacting parts is strongly influenced by the tribological properties of the system . adhesion and friction play a critical role in determining the operational reliability and lifetimes of such systems @xcite . one of the main causes of adhesion and friction problems of mems is the presence of a thin water film present on most surfaces . this is particularly true of silicon , the most popular of all the mems materials . freshly exposed silicon rapidly oxidizes in air forming a nm thick hydrophilic oxide layer , which can ultimately dictate the performance of the entire device . thus , it is important to thoroughly understand the adhesion and friction properties of contacting surfaces down to the micro and nanoscales . this knowledge will facilitate the fabrication of next generation micro and nanodevices with tailored tribological characteristics . previous investigations have examined different aspects of adhesion and friction properties of materials on the micro and nanoscale . the microfrictional properties of a tribosystem consisting of a 1 mm sapphire sphere sliding against a flat silicon sample was investigated from air to ultrahigh vacuum ( uhv ) @xcite . this study showed that friction initially decreased exponentially as a function of the pump - down time . later , as more water desorbed from the tribocontact distinct peaks with higher friction were seen in the friction versus pump - down time curve . these friction peaks were characterized by the occurrence of distinct stick / slip friction processes and regions between them that were absent from such effects . it was postulated that the peaks with stick / slip friction was caused by the solid - like friction of thin ordered water films confined between the ball and sample . it was further postulated that friction peaks occur whenever an ordering effect occurs due to the formation of a complete monolayer , which results in solid - like friction . in the transition period between complete monolayers , the molecules were judged to be not ordered . in this regime , friction was characterized only by small fluctuations and was liquid - like . another group determined the coefficient of friction for two flat silicon samples @xcite in a simple experiment . the lower sample was tilted until the upper sample just begins to slide . by recording the tilt angle , @xmath0 , when sliding occurred the coefficient of friction , @xmath1 was calculated from the relationship : @xmath2 . the experiment showed that , when water coverage on the surface is reduced due to desorption in vacuum , friction is first reduced due to decreasing capillarity and is then followed by increased friction . similar experimental results were obtained in experiments on single - crystalline ni(100 ) @xcite where friction was measured for decreasing thickness of an ethanol film in uhv . smooth sliding at a coverage of about 300 monolayers changed into stick / slip sliding when the ethanol film was reduced to around a monolayer . this was then followed by strong sticking in the submonolayer range . the contact pressures in the above experiments were of the order of tens or at most hundreds of mpa , which is not extremely high compared to most macro and even nano contact situations . due to this characteristic , a liquid layer is present at all times on the surfaces between contacting parts and the confined liquid films are responsible for causing stick / slip effects etc . in contrast , in a typical sfm contact the contact area is much smaller ( @xmath320 nm@xmath4 ) . with a normal force on the order of 50 nn , the contact pressure comes out to be in the range of several gpa . this unique contact feature results in a tribological behavior that is different from the microfrictional studies discussed above . in a nanotribological study , friction as a function of the sliding velocity , relative humidity and wettability was examined by riedo et al . @xcite . in this study , a linear dependence of friction force vs. logarithm of sliding velocity was observed . the slope for hydrophilic surfaces was determined to be negative and for higher sliding velocities the friction force decreased ; the slope was reduced at lower humidities . the results were interpreted to indicate that the influence of capillary force is reduced with decreasing humidity and at higher sliding velocities . it was also shown that the slope for hydrophobic surfaces is positive and does not change with varying humidity . a nanotribological investigation , in which the function of friction force and water film thickness vs. residual gas pressure during pump down in a vacuum chamber , was conducted by opitz et al . @xcite . in this report , different friction and contact regimes were identified . the water film thickness for hydrophilic native oxide covered silicon samples , measured by distance dynamic force spectroscopy , was found to reduce from about 2.6 nm to 0.7 nm with decreasing residual gas pressure in three distinct regimes . further measurements showed the complete removal of water due to the combined effect of the vacuum environment and friction - induced desorption . the measured friction force also showed three regimes . it was seen that the reduction of the friction force corresponds to decrease of the water film thickness . this work extends the knowledge gained in the above mentioned study @xcite in that it examines the dependence of friction on velocity and on an externally applied normal load and uses this information to derive nanofriction mechanisms and basic nanocontact properties of the tribosystem . a continually decreasing water film thickness was produced by inserting the sample in a vacuum system and pumping it down from ambient pressure to 10@xmath5 mbar . hydrophilic , native oxide covered silicon was used as sample and counterbody . distinct regimes of friction versus the applied normal load and the sliding velocity could be identified . rectangular pieces of p - doped ( 1 - 10 @xmath6 cm , b - doped ) native oxide covered si(100 ) were used as samples . these had an sfm determined rms ( root mean square ) roughness of about 0.1 nm measured over a [email protected] @xmath8m@xmath4 area . due to the presence of a native oxide on top of the si(100 ) surface , the samples are oh - terminated @xcite and hydrophilic @xcite . the hydrophilicity was also separately confirmed by contact angle measurements ( not shown ) . prior to insertion into the uhv chamber , the samples were sequentially cleaned using ultrasonic assistance in isopropanol and methanol for five minutes each . afterwards , the samples were thoroughly rinsed with bi - distilled water . the oxide film measured on such samples has a thickness of 1.3 nm as determined by x - ray induced photoelectron spectroscopy ( xps ) . a uhv vacuum sfm / stm ( omicron nanotechnology gmbh ) system was operated in contact sfm mode to measure friction between a sharp silicon tip and a planar si(100 ) sample . measurements were performed at various stages during evacuation of the system , from air to 10@xmath5 mbar , using a combination of turbo molecular , ion and titanium sublimation pumps . starting from air , upon pump - down the chamber pressure decreased rapidly between 10@xmath9 and 10@xmath10 mbar . due to the fact that at this stage thermodynamic equilibrium does not exist , no measurements for the velocity and normal force characteristic could be performed . friction measurements using contact sfm were performed in lateral mode registering the friction hysteresis at different applied normal loads with sliding velocity of 300 nm / s and at different sliding velocities with an applied normal load of 60 nn . for each friction measurement the tip was scanned 300 nm in the forward and backward direction and the friction was determined by taking one half of the friction hysteresis curve @xcite . to attain a high lateral force resolution , single beam si cantilevers with a length of 440 @xmath8 m were chosen instead of triangular cantilevers that are typically used for topographical measurements . these single beam cantilevers , with a normal spring constant of 0.11 n / m , and a manufacture - quoted tip radius that was smaller than 15 nm , were used as received , i.e. in an oxidized ( hydrophilic ) state . calibration of the normal and lateral forces was achieved by following the procedure developed by schwarz et al . @xcite , which consists of using the geometrical dimensions of the tip and determining the magnitude of tip deflection as a function of the applied load . figure [ fig:1 ] shows the friction force measured as a function of the residual gas pressure from air down to 10@xmath5 mbar . in this case , the silicon sample was cleaned as described above and then thoroughly rinsed with water before inserting it into the vacuum system . after this preparation process the hydrophilic sample is covered by a water with a thickness of 2 nm . initially , a high friction force was measured ( fig . [ fig:1 ] ) . afterwards , during pump - down , friction decreased from range ( a ) to range ( b ) in fig . [ fig:1 ] to half the initial value . this is due to water desorption when the residual gas pressure is decreased . however , a subsequent reduction in the friction also occurred at lower pressures . it is suggested that this last step ( marked ( c ) in fig . [ fig:1 ] ) to lower friction values occurs due to the combined effect of the vacuum and friction - induced desorption of water from about 0.7 nm thickness to almost complete water removal . the resulting contact at this stage is then a direct solid - solid contact without the presence of an additional water film . the friction force vs. sliding velocity dependence , characteristic for the three ranges shown in fig . [ fig:1 ] , is depicted in fig . [ fig:2 ] . the sliding velocity is plotted in logarithmic scale . in this plot a linear dependence is visible . the equation used to fit this measurement is given as @xcite @xmath11 here are @xmath12 the intercept and @xmath13 the slope . the constant @xmath14 was fixed to 1 nm / s to keep the units correct . figure [ fig:3 ] shows the changing of the slope @xmath13 as a function of the residual gas pressure . like in figure [ fig:1 ] , here also the three distinct regions : ( a ) , ( b ) and ( c ) can be easily discerned . the slope changes from high negative values for regime ( a ) to positive values in regime ( b ) and ends at comparatively small positive values for regime ( c ) . the boundaries between these regimes are the same like in the friction vs. residual gas pressure measurements ( see fig . [ fig:1 ] ) . bowden and tabor @xcite described the dependence of friction force on the applied normal force with respect to the contact area . later , maugis @xcite derived a value for the contact area from the elastic theory of solids and from rupture mechanics . in this work a parameter @xmath15 is defined whose magnitude indicates which borderline case of the general contact model should be applied . it is defined as @xmath16 where @xmath17 is the surface energy , @xmath18 the interaction length in the dugdale interaction model , @xmath19 the tip radius and @xmath20 the reduced elastic modulus . using the values @xmath21 nm @xcite , @xmath22 mj@xmath23m@xmath24 @xcite , @xcite and assuming that the interatomic distance , @xmath18 is 0.5 nm the calculation for @xmath15 gives 0.07 . when @xmath15 is smaller than 0.1 the dmt - model @xcite for the friction force @xmath25 can be used : @xmath26 here @xmath27 is the shear stress in the friction contact , @xmath17 the surface energy and @xmath28 the adhesion force in the dmt - model . the normal force in this model is the sum of applied normal load @xmath29 and the adhesion force @xmath30 . the dependence of friction force on the applied normal load is shown in figure [ fig:4 ] for the three distinct friction regimes ( as was the case in previous figures ) . the intercept on the load axis as well as the values for the friction forces change from region ( a ) to ( c ) . to extract more information the measured curves were fitted using equ . [ eq:3 ] and the surface energy and the shear stress were then evaluated ( see fig . [ fig:5 ] ) . again a value of @xmath31 gpa was used and @xmath21 nm was taken as the tip radius . this is larger than the manufacture - quoted tip radius ( less than 15 nm ) because our studies indicate that the tip gets blunter during measurement @xcite . the surface energy , @xmath17 , and the shear stress , @xmath27 , ( fig . [ fig:5 ] ) reduce in two steps during pump - down . the three distinct pressure regimes , however , are still visible . the dependence of friction on the sliding velocity and applied normal load for a silicon tip sliding against a silicon sample changes considerably from wet at air pressure to very dry in ultrahigh vacuum . the nanofriction between two hydrophilic surfaces ( fig . [ fig:1 ] ) is reduced in three regimes . the slope in the friction force vs. logarithm of sliding velocity curves ( fig . [ fig:3 ] ) describes the friction behavior of the nanotribological contact . the negative slope in regime ( a ) shows the domination of capillary forces . the friction force is determined by the number of capillary bridges that form in the nanocontact between tip and sample @xcite . the number of capillary necks and the influence of capillary force are reduced at higher sliding velocities assuming an activation process for the time to build the water bridges @xcite . when the water film thickness is above 2 nm @xcite , for tribological purposes water can be considered as bulk water . during pump - down , the slope changes to positive values after reduction of the residual gas pressure ( and with that the water partial pressure ) . at this stage , bulk water desorbs when the water partial pressure of ambient is lower then the water vapor pressure in regime ( b ) . the residual water film remaining on the sample surface , has a film thickness of about 0.7 nm and correlates to a film composed of 2 ice - like water bilayers @xcite . in this regime , ordering effects of the ice - like bilayers strongly influence friction . we speculate that friction arises from the `` pushing aside '' of these bilayers by the sliding tip , which has always some solid - solid contact present due to the high applied normal load ( pressure ) in this study @xcite . these ordering effects result in the positive slope observed in fig . [ fig:2 ] . the decreasing of the slope in region ( c ) is attributed to the complete removal of water due to desorption . a similar small positive slope was also measured on hydrophobic @xcite as well as on atomically flat @xcite surfaces . thus , our measurements demonstrate the transition of friction from a wet , hydrophilic surface to a dry hydrophilic one . from these results one can deduce that the characteristics of a dry hydrophilic surface are comparable to those of hydrophobic surfaces . the frictional behavior derived from the nanocontact conditions is defined by the structure of the water film present on the surface . the water film in regime ( a ) is bulk water on top and an electrochemical double layer between this bulk water and silicon oxide surface @xcite . a comparison between bulk water in regime ( a ) and ordered water in ( b ) ( figure [ fig:2 ] ) shows that , when an extrapolation of the friction force vs. sliding velocity curves in regime ( a ) and ( b ) is made to higher velocities , they cross at a velocity of about 2 mm / s . this result suggests that at this velocity the influence of capillary force vanishes and friction is dominated by ordering effects of the thin water layer . in other words , this means that the slope of the friction vs. logarithm of sliding velocity should reverse at velocities in the millimeter per second range for the examined contact parameters . however , since the highest achievable sliding velocity in the used device is about 14 @xmath8 m/s , experimental verification of this phenomenon was not possible . according to the dmt - model @xcite , the surface energy relates to the adhesion force as @xmath32 . following this equation , adhesion increases for higher surface energies . in our experiments , the surface energy undergoes a reduction in 3 regimes ( fig . [ fig:5 ] ) . the difference of surface energy between regime ( a ) and ( b ) is about 122 mj@xmath23m@xmath24 and is in the order of magnitude of the surface tension of bulk water ( @xmath33 mj @xmath23m@xmath24 ) . this confirms that bulk water desorbs during the transition from regime ( a ) to ( b ) . the lowest value of surface energy is reached at pressures lower then 10@xmath34 mbar , where it has a value of 21 mj@xmath23m@xmath24 . this value is comparable to measurements determined by adhesion occurring in a micro cantilever array @xcite . another quantity , the shear stress , describes the lateral force per unit area for two moving parts in contact . in our experiments , the shear stress changes in two steps ( fig . [ fig:5 ] ) . the highest shear stress is measured at high water film thicknesses in regime ( a ) . at this stage , the effects of solid - solid contact , ordering of the electrochemical double layer and bulk water , resist the motion of the tip . the removal of the bulk water layer shows decreasing of shear stress in regime ( b ) . at this stage , interactions due to the solid - solid contact and the ordering effects of the two ice - like water bilayers dictate the frictional properties . the lowest shear force ( visible in regime ( c ) ) is present when only solid - solid contact occurs . it should be noted that at the applied normal load in this study , solid - solid contact also occurs in the regime where ordering effects of thin water double layers are present . also , these two effects occur in the capillary force regime where bulk water is present on the surface . the friction force was examined as a function of the residual gas pressure , sliding velocity and applied normal load using contact sfm . the friction force on hydrophilic silicon undergoes changes during pump - down in a vacuum system from air to 10@xmath5 mbar . three distinct friction regimes with different nanofriction mechanisms and nanocontact characteristics , were identified . these regimes are shown schematically in figure [ fig:6 ] . first , the capillary force dominates . later , during pump - down , most of the water is removed by vacuum desorption . the remaining water is now an ordered water double layer and the effects of the ordered water layer dominate friction . after complete water desorption due to the combined effects of the vacuum and friction - induced desorption , only solid - solid contact remains , which exhibits the lowest friction of all the three friction regimes . in this study , the water present on the tip and sample surface never acts as a lubricant . liu , h. , s. i .- u . ahmed , and m. scherge : 2001 , ` microtribological properties of silicon and silicon coated with diamond like carbon , octadecyltrichlorosilan and stearic acid cadmium salt films : a comparative study ' . , 135 . tanner , d. m. , j. a. walraven , l. w. irwin , m. t. dugger , n. f. smith , w. p. eaton , w. m. miller , and s. l. miller : 1999 , ` the effect of humidity on the reliability of a surface micromachined microengine ' . p. 189 . friction force vs. sliding velocity for the three friction regimes identified in fig . [ fig:1 ] . for every friction regime a representative measurement the solid lines are fits using equation [ eq:1 ] . slope @xmath13 from equation [ eq:1 ] vs. the residual gas pressure . the three different regimes from fig . [ fig:1 ] are marked using the vertical grey lines . they correspond well with the change in the slope @xmath13 . the solid line connecting the data points has been drawn to guide the eye . friction force vs. applied normal load for the three friction regimes from fig . [ fig:1 ] . a representative measurement is shown for each regime . the solid lines are fits using equation [ eq:3 ] . a zoomed diagram is shown for regime ( c ) for better visibility . surface energy and shear stress as determined from equation [ eq:3 ] vs. the residual gas pressure . the three different friction regimes from fig . [ fig:1 ] are marked using the vertical grey lines . the solid line connecting the data points has been drawn to guide the eye . schematic model of water coverage in various coverage regimes determined by the friction behavior . while there is only the solid - solid interaction in regime ( c ) , this interaction along with ordering effects of water are both present in regime ( b ) and , similarly , the solid - solid interaction and ordering effects are also present in regime ( a ) .	this paper examines friction as a function of the sliding velocity and applied normal load from air to uhv in a scanning force microscope ( sfm ) experiment in which a sharp silicon tip slides against a flat si(100 ) sample . under ambient conditions , both surfaces are covered by a native oxide , which is hydrophilic . during pump - down in the vacuum chamber housing the sfm , the behavior of friction as a function of the applied normal load and the sliding velocity undergoes a change . by analyzing these changes it is possible to identify three distinct friction regimes with corresponding contact properties : ( a ) friction dominated by the additional normal forces induced by capillarity due to the presence of thick water films , ( b ) higher drag force from ordering effects present in thin water layers and ( c ) low friction due to direct solid - solid contact for the sample with the counterbody . depending on environmental conditions and the applied normal load , all three mechanisms may be present at one time . their individual contributions can be identified by investigating the dependence of friction on the applied normal load as well as on the sliding velocity in different pressure regimes , thus providing information about nanoscale friction mechanisms .
when a scale invariant free field is perturbed by an interaction , the scaling symmetry is in general broken . in the case of the free massless scalar field in 4-dimensional minkowski space , this `` conformal anomaly '' is well known : the renormalization of loop diagrams requires the introduction of a scale parameter which breaks scale invariance . using the non - uniqueness of renormalization , the best one can reach is `` almost homogeneous scaling '' , i.e. the breaking terms for the scaling @xmath9 are proportional to some power of @xmath10 . ( for a systematic treatment in the framework of causal perturbation theory see @xcite . ) in this paper , we want to address the analogous issue for scale invariant _ generalized free fields _ ( free fields with non - canonical scaling dimension , see below ) . such fields naturally arise as boundary limits of klein - gordon fields on ads @xcite . the basic question is : * _ is it possible to construct _ scale invariant interacting fields _ ( admitting for anomalous dimensions ) ( ^l)_l(x)=+o ( ) as perturbative expansions around wick powers @xmath11 of scale invariant generalized free fields @xmath0 @xcite ? _ ( @xmath12 denotes the interaction density and @xmath13 the coupling constant . ) perturbation theory around a generalized free field ( in minkowski space ) suffers from a huge arbitrariness which is due to renormalization , as we point out in sect . [ sec : procedure ] . on the other hand , the requirement of scale invariance is very restrictive . in important cases ( which we do not want to exclude ) it can not be fulfilled even for tree diagrams ( sect . [ sec : comparison ] ) . namely , the propagator needs a nontrivial renormalization if the scaling dimension @xmath14 is @xmath15 in four dimensions ( @xmath16 in @xmath1 dimensions ) , and for integer @xmath14 a breaking of scale invariance can not be avoided . we propose here a method to circumvent these difficulties and construct perturbatively interacting fields with unbroken conformal symmetry , by taking advantage of the ads - cft correspondence . viewing a conformally covariant field on minkowski space - time as a boundary limit of an ads covariant field on anti - desitter space - time @xcite , an ads invariant renormalization in the bulk guarantees an anomaly free conformal symmetry of the boundary field , provided the boundary limit exists . in this way , the ads - cft correspondence turns out to be a useful tool also when one is only interested in cft in minkowski spacetime . in @xcite and @xcite it was shown that the boundary limit @xmath17 of @xmath18 such that @xmath19 lies on the hyperboloid @xmath20 w.r.t . to the metric of signature @xmath21 in the ambient space @xmath22 . the @xmath23 metric is the induced one : @xmath24 , see e.g. @xcite . ] of the scalar klein - gordon field @xmath25 of mass @xmath26 on @xmath27-dimensional ads is a generalized free field @xmath28 with scaling dimension [ dim ] = _ + = d2+,(= ) , see sect . [ sec : procedure ] . the corresponding boundary limit of the free wick powers @xmath29 yields fields @xmath30 which have scaling dimensions @xmath31 . notice that in the witten model @xcite of maldacena s conjectured ads - cft correspondence @xcite , one studies instead the `` dual '' field with boundary conditions corresponding to @xmath32 , which is coupled to the sources in a `` dual '' way . however , it was shown in @xcite that the dual coupling modifies the relevant bulk propagator by a correction term in such a way , that the full propagator becomes that of the above klein - gordon field , and the unrenormalized perturbative expansion of the dually coupled boundary field is formally equivalent to the boundary limit of the bulk field @xmath25 with the same interaction . ( the same nontrivial features , that are of representation theoretic nature , were established for the propagators of tensor fields of any rank @xcite . ) regarding the generalized free field as a limit of a canonical free field on ads , the task is to extend this relation to the renormalized interacting fields . hence , we first construct the interacting ads fields w_l(z , x)= + o ( ) for polynomial interactions @xmath33 in sect . [ sec : phi ] and sect . [ sec : phi2 ] , using standard renormalization methods of causal perturbation theory ( reviewed in sect . [ sec : causalpt ] and [ sec : renorm ] ) . at this stage , the non - uniqueness of the renormalization can be classified by the usual short distance power counting @xcite , and the propagator is unique and ads - invariant , hence the ads symmetry is fully preserved . then , the essential step is to investigate the existence of a boundary limit [ wkl ] w_l(x ) = _ z0 z^-^w_l w_l(x ) in the renormalized theory . here , we admit for anomalous dimensions , i.e. , @xmath34 . if this limit exists , we prove that it inherits the ads symmetry of the bulk as an exact ( unbroken ) conformal symmetry ( sect . [ sec : conf ] ) . our main result is that the boundary limit does exist , for typical polynomial interactions , for the interacting field ( sect . [ sec : phi ] ) and for composite fields ( sect . [ sec : phi2 ] ) , due to nontrivial cancellations within the renormalized one - loop distributions taking place in the limit . although the actual computations are `` hidden '' in apps . [ app - pf ] and [ app : limit ] , these cancellations constitute the essential mechanism to allow the passage to the boundary . in order to establish this result , along the way we develop a `` universal '' formula ( lemma b.1 in app . [ app : lemma ] ) that controls the asymptotic behaviour near the boundary of a large class of typical interactions and diagrams . thus , the above posed question gets an affirmative answer for those interactions @xmath35 $ ] of the conformal field which are `` induced '' by the corresponding polynomial ads interaction @xmath36 $ ] ( as indicated by retaining the subscript @xmath37 in also for the boundary field ) . this means @xcite that d^dx l[(x ) ] = dzd^dxl[(z , x ) ] , hence the cft interaction density l[(x ) ] = dz l[(z , x)]= l[_h_z(x)][action ] arises as the @xmath38-integral over @xmath39 $ ] where @xmath40 is the ads field @xmath25 re - expressed as a family of boundary generalized free fields belonging to the borchers class of @xmath0 ( @xcite , see sect . [ sec : freefields ] ) . we point out that , due to the integration in , the interaction vertices `` remain in the bulk '' . in this sense , the situation is converse to rhl s reconstruction @xcite of an ads field from an interacting conformal field where the ads interaction is restricted to the boundary ( namely , the ads field in @xcite satisfies the free field equation in the bulk ) . it is an essential aspect of our approach that , while the general principles of renormalization are the same , the detailed implementation of the rules differ in the bulk and on the boundary . in order to exhibit the methodic difference which allows the renormalization in the bulk to preserve the symmetry that is necessarily broken by renormalization on the boundary , we compare both approaches in sect . [ sec : countex ] with a flat space toy model , where this difference is much more transparent . let us recall @xcite how the klein - gordon field on @xmath27-dimensional anti - desitter space and generalized free fields on @xmath1-dimensional minkowski space can be represented in terms of the same creation and annihilation operators , and hence as field operators on the same hilbert space . the free klein - gordon field @xmath2 of mass @xmath26 on ads can be expressed as ( z , x ) = 12 z^d2 _ 0^dm^2 j_(mz)_m(x ) , [ phi : kg ] where @xmath41 is a massive free boundary field given by _ m(x)_k_00d^dk(k^2-m^2 ) . [ ads : kg ] the parameter @xmath42 is related to the mass by @xmath43 . the functions @xmath44 are the plane - wave solutions to the klein - gordon equation on ads , where the laplacian is [ laplace ] _ x = -z^1+d_z z^1-d_z + z^2_x , and @xmath45 ( @xmath46 ) are creation and annihilation operators normalized as [ a(k),a^+(k ) ] = ( 2)^-(d-1)^d(k - k ) , = 0 = [ a^+,a^+ ] , in the fock space @xmath47 over the continuous mass 1-particle space @xmath48 . in this hilbert space , the fields _ d^dk h(k^2 ) [ gff : h ] ( with @xmath49 any sufficiently smooth polynomially bounded real function on @xmath50 ) are local and poincar covariant generalized free scalar fields in @xmath1-dimensional minkowski space with kllen - lehmann measure @xmath51 . thus , @xmath2 may be written as ( z , x)=_h_z(x ) h_z(m^2 ) z^d2 j_(zm ) . [ phi = varphi_h ] taking the boundary limit , we get @xcite : _ z0 z^- ( z , x ) = ( x ) [ boundlim ] with was denoted @xmath52 in @xcite , whereas @xmath53 with @xmath54 was denoted @xmath0 . ] ( x)=c__v_+ d^dk ( k^2)^ [ a(k ) e^-ikx + a^+(k)e^ikx ] , + d2 , c _ , [ gff ] i.e. , @xmath55 with @xmath56 . its kllen - lehmann measure being a homogeneous function of the mass : d(m^2)=c_^2 m^2 dm^2 , [ weightf ] the boundary field @xmath0 is scale invariant : u()(x)u()^=^ ( x ) , [ scaling ] and in fact transforms like a conformal scalar field under the representation of the ads symmetry group on the fock space of the ads klein - gordon field @xmath2 . the boundary limit can obviously be generalized to arbitrary wick polynomials @xmath57 , w(x)=_z0 z^-lw(z , x)= [ boundlim - w ] which have scaling dimension @xmath58 ( where @xmath59 is a multi - index ) . the aim of this paper is to investigate causal perturbation theory @xcite around the generalized free field ( and its wick polynomials ) . causal perturbation theory proceeds @xcite by defining , for each wick polynomial @xmath60 of free fields @xmath2 , the interacting field @xmath61 as formal expansion in wick products of the free field @xmath2 with distributional coefficients . this expansion is obtained as the exponential series of retarded products of @xmath60 with the interaction @xmath62 , where the retarded products are operator - valued distributions . they are determined recursively ( by the postulated causal properties of the interacting fields ) _ at non - coinciding points only _ ; the renormalization of the perturbative expansion consists in the _ extension of these distributions to coinciding points_. `` renormalization conditions '' ( covariance , ward identities , ) serve to reduce the arbitrariness in the extension , and the main problem is to decide whether all desirable renormalization conditions can be fulfilled at the same time , with a finite number of free parameters remaining . this program is performed with the interaction being cut off in space and time by means of a space - time dependent coupling constant @xmath63 . it then remains to control the _ adiabatic limit _ of removing the cutoff , @xmath64 . this limit is in general plagued by infrared problems ; it is , however , possible to define the _ algebraic adiabatic limit _ @xcite , i.e. , the local field algebras @xmath65 in arbitrary bounded space - time regions @xmath66 , without infrared problems as long as the construction of the interacting vacuum state is postponed . causal perturbation theory around a generalized free field is , however , problematic for the following reason . to construct the general solution for the perturbative @xmath67-matrix one has to use the wick expansion formula for time - ordered or retarded products ( also called the `` causal wick expansion '' ) @xcite . for simplicitly , let us discuss here the ordinary wick expansion formula , which for mass shell free fields is [ wick ] = + _ r_1, ,r_n_i=1^n ( c k_i + r_i ) ( , ) . for generalized free fields , the lagrangean can be any field relatively local w.r.t . the generalized free field , i.e. , any element of its borchers class . the borchers class contains at least the `` generalized wick polynomials '' @xcite [ gwp ] ( ) _ h(x ) = + _ v_+ d^dk_1 _ v_+ d^dk_l h(k_1 ^ 2, ,k_l^2 ) where @xmath68 is any symmetric and sufficiently regular function . let us choose a lagrangean @xmath69 with an arbitrary function @xmath70 . it is then easy to see , that the wick expansion of , say , @xmath71 with @xmath72 does not factorize as in , but rather contains terms of the form [ gwick ] _ v_+ d^dk_1 d^dk_3 h(x - y;k_1 ^ 2,k_2 ^ 2,k_3 ^ 2 ) , where h(x - y;k_1 ^ 2,k_2 ^ 2,k_3 ^ 2 ) = _ v_+dq e^-iq(x - y)h(q^2 ) h(q^2,k_1 ^ 2,k_2 ^ 2,k_3 ^ 2 ) . because the dependence of this function on @xmath73 and on @xmath74 is entangled in a nontrivial manner , the numerical distribution can not be separated from the operator - valued distribution as in ( unless @xmath75 happens to be a factorizing function ) . interpreting as an operator product expansion , reveals a characteristic feature of the theory of generalized free fields : performing first the @xmath76-integrations , the subsequent @xmath77-integration may be interpreted as a `` continuous sum '' over generalized wick products . more importantly , however , the failure of separation as in would require more refined methods to establish the existence of a renormalization , than the standard methods of causal perturbation theory , which proceeds by renormalizing only the numerical distributions ( see below ) . let us contrast the general case to the case when the interaction is induced by a local interaction on ads @xcite as described in the introduction , i.e. , when the conformal field @xmath0 arises as the boundary limit of a canonical ads field @xmath2 with interaction @xmath37 . the lagrangean @xmath12 given by with , say , @xmath78 on ads is @xmath79 with h(k_1 ^ 2, ,k_4 ^ 2)=dzz^-d-1h_z(k_i^2 ) , i.e. , @xmath75 is a @xmath38-integral over factorizing functions ; one can therefore reorganize the continuous ope as a @xmath38-integral over wick products of the distinguished fields @xmath40 as in , rather than generalized wick products as in . this fact seems to reduce the renormalization ambiguity drastically , since the freedom is only in the choice of suitable weight functions in @xmath38 . whether a conformally covariant renormalization of the ope of perturbed boundary fields is possible , would require a nontrivial analysis . this is the reason why we propose to work instead with the `` bulk approach '' mentioned before , using the correspondence and ; i.e. , we first construct the perturbative interacting fields on @xmath27-dimensional anti - desitter space @xcite , and then study their boundary limit . we shall see that conformal covariance can be maintained on the boundary because ads covariance can be maintained in the bulk . the issue therefore has been shifted to the existence of the limit . it will be illustrated in sect . [ sec : countex ] , why this indirect approach gives different results than the direct approach perturbing generalized free fields on the boundary . in @xcite and @xcite perturbative interacting fields have been constructed on an arbitrary globally hyperbolic curved spacetime @xmath80 for localized interactions @xmath81 , i.e. , the interaction @xmath82 is switched on by @xmath83 . the anti - desitter spacetime is not itself globally hyperbolic , but its covering is conformally equivalent to a @xmath84 quotient of a globally hyperbolic space - time @xcite . in this way , the lack of global hyperbolicity can be circumvented in terms of boundary conditions `` at infinity '' ( @xmath85 ) . if one wants to take the boundary limit , one obviously must not cut off the interaction on the boundary of ads , hence we must perform a `` partial adiabatic limit '' which puts the switching function @xmath86 to be 1 for @xmath87 ( a compact region @xmath88 ) and @xmath85 . it can be easily seen that the conclusion of @xcite , i.e. , the independence of the algebraic adiabatic limit on the details of the switching function outside the compact region of interest , holds also true for the partial adiabatic limit . we may therefore assume that the switching function factorizes as g(z , x)=(z)g(x ) , g _ [ 0,a ] k= [ g = cg ] with @xmath89 and @xmath90}\equiv 1 $ ] for some @xmath91 . in addition @xmath92 and @xmath93 are smooth , @xmath94 is compact and the support of @xmath95 is bounded for @xmath96 . since the support of such functions @xmath97 are not compact in ads , there may in principle be ir problems associated with the partial adiabatic limit ; but our explicit calculations in sect . 3 show that these do not appear in the relevant examples . the ( partial ) algebraic adiabatic limit does not depend on the details of the functions @xmath92 and @xmath93 , provided @xmath98 is sufficiently large . in practice , we proceed as follows : given a wick monomial @xmath99 in the generalized free field @xmath0 and its derivatives , we first replace @xmath28 by the ads field @xmath25 ( whose boundary limit is @xmath28 ) , and construct the interacting ads field @xmath100 associated with the corresponding wick monomial @xmath60 in @xmath2 and its derivatives . then we define the interacting field @xmath101 in @xmath102 as boundary limit of the interacting field @xmath100 on @xmath23 , provided this limit exists : w_l(x)=_z0 z^-^w_lw_l(z , x ) , [ w = limw ] where ^w_l = l+_n=1^^n ( ^w_l)^(n ) . [ delta^w ] the deformation @xmath103 ( i.e. , the sequence of coefficients @xmath104 , @xmath105 ) is determined by the requirement that the limit exists . remark : * in ordinary perturbative qft the anomalous dimension is the deviation of the scaling dimension of an ( interacting ) quantum field @xmath106 from the scaling dimension of the corresponding ( interacting ) classical field . for generalized free fields there is no obvious classical counterpart . instead , we call `` anomalous dimension of @xmath107 '' the deformation of the scaling dimension _ due to the interaction_. in contrast to ordinary perturbative qft , it does not come from the breaking of scale invariance in the renormalization of loop diagrams ( we maintain the ads - symmetry in the renormalization ) . instead its appearance is enforced by the existence of the boundary limit . in causal perturbation theory on ads , @xmath108 is given by @xcite [ w_gl ] w_l(x)=_n=0^ ( _ r=1^n_0^ ( z_r ) d^dx_rg(x_r ) ) + r_n,1(l(x_1), ,l(x_n);w(x ) ) where @xmath109 and @xmath110 . the unrenormalized retarded products @xmath111 are determined as distributions at non - coinciding points @xmath112 . the result is @xcite [ r : unren ] r_n,1(l(x_1), l(x_n);w(x ) ) = + ( -i)^n n ! s where @xmath113 means symmetrization in @xmath114 . now let @xmath115 and @xmath116 . then , using and , the retarded product may be rewritten as [ r(varphi_h ) ] & & r_n,1 ( , , ; ) = + & & = ( -i)^n n ! s ( x^0>x^0_n > x_n-1 > > x^0_1 ) + & & _ r=1^n_s=1^kdm^2_rsj_(m_rsz_r ) _ j=1^l dm_j^2j_(m_jz ) + & & . we emphasize that writing as in the left - hand side of is misleading : _ it is not a retarded product in minkowski space _ , but in @xmath23 , defined with respect to the causal structure in @xmath23 . in particular , the problem with the causal wick expansion for generalized free fields mentioned before , is absent , and its correct _ definition _ is the right - hand side of . moreover , renormalization is needed for coinciding ads points @xmath117 only , and not on the whole submanifold @xmath118 , as will be discussed in the next subsection . in the sequel , we shall be mainly concerned with special cases of the type r_1,1(;(x ) ) = k i(x;x_1)(x^0-x^0_1 ) and [ typical ] r_1,1 ( ; ) = 2ki(x;x_1)(x^0-x^0_1 ) + + + k(k-1 ) i ( ^+(x;x_1)^2-^+(x_1;x)^2)(x^0-x^0_1 ) where @xmath119 is the scalar 2-point function , and @xmath120\omega)$ ] the commutator function . the expressions are not defined as distributions at coinciding points , due to the time - ordering @xmath121 functions . the problem of renormalization is thus the extension of the retarded products to distributions @xmath122 on @xmath123 . by the recursive construction principle underlying causal perturbation theory , once this has been achieved for @xmath124 ( @xmath125 ) , then @xmath111 is already determined everywhere outside the _ total _ diagonal _ n+1\{(x_1, ... ,x_n;x)\| x_j = xj=1, ... ,n}. [ diagonal ] renormalization at @xmath126th order is thus reduced to the extension of the distributions @xmath111 from @xmath127 to @xmath123 . applying the recursion as indicated , gives rise to a diagrammatic expansion of @xmath111 in terms of wick products with propagators and numerical distributions @xmath128 ( @xmath129 ) as coefficients , as in . the latter are the vacuum expectation values of operator - valued distributions ( with field arguments of possibly lower order ) . e.g. , for @xmath130 and @xmath116 , there arises the `` fish diagram '' ( fig . 1 ) as the coefficient of @xmath131 to first order in @xmath13 . arising in first order perturbation theory for the interacting field @xmath132 with interaction @xmath33 . the diagram symbolizes the distribution @xmath133 or the corresponding unrenormalized expression @xmath134 ( given by or resp . , appearing in the second line of ) . renormalization is done in terms of the _ numerical _ distributions @xmath135 , by extending them to distributions @xmath136 on @xmath137 . ( for an example in flat space , see sect . [ sec : countex ] ) we shall see , however , that the @xmath17 behaviour of the renormalized operator - valued distributions @xmath111 on ads is in general not the same as that of the numerical distributions @xmath136 ; thus the existence of the limit has to be studied for the operator - valued distribution @xmath111 . for a rigorous and complete definition of the retarded products @xmath111 we refer to the renormalization axioms given in @xcite when @xmath138 and @xmath116 . ] , with appropriate modifications due to the curvature of @xmath23 @xcite . in particular , the renormalization should not increase the scaling degree of a distribution @xcite , which controls the `` strength of the uv singularity '' : the scaling degree is defined in flat space by _ y(f(;x ) ) = inf \ { _ 0^f(x+y;x)=0 } , where the limit is meant as a distribution in @xmath139 ; in curved spacetime , @xmath140 is taken in the tangent space and the argument @xmath141 has to be replaced by the geodesic exponential @xmath142 . moreover , the renormalization conditions of translation invariance and @xmath143-covariance are replaced by ads - invariance ( group @xmath144 ) . the expression is obviously ads - invariant , so the problem consists in preservation of this symmetry upon renormalization . since we construct the interacting field on ads , @xmath145 needs to be renormalized at @xmath146 @xmath147 only , while at @xmath148 for all @xmath76 , @xmath149 for some @xmath76 , it is already defined by the recursion . this fact is responsible for a drastic reduction of renormalization ambiguities in the ads approach , as compared to renormalization of generalized free fields on minkowski space . the renormalization freedom is further reduced by _ requiring _ the existence of a boundary limit as a renormalization condition . we shall see in some typical examples ( sect . [ sec : phi ] and sect . [ sec : phi2 ] ) that this condition may require a `` field mixing '' , i.e. , perturbative corrections of an interacting wick monomial by @xmath150 times other wick monomials , in order to cancel perturbative contributions of different scaling dimensions . we shall show in the next subsection that for @xmath138 ( no derivatives ) , the ads covariant renormalization of @xmath108 ensures conformal covariance of its boundary limit w_l(x ) = _ z0 z^-^w_lw_l(z , x ) [ boundarylimit ] provided this limit exists , with a suitable ( coupling dependent ) scale dimension @xmath151 . then we shall illustrate the difference between renormalization on ads and renormalization on the minkowski boundary by a flat space model which avoids the technical complications of the curvature . in sect . 3 , we shall address the renormalizability on ads and the existence of the boundary limit with some case studies . in this subsection we assume that for a polynomial interaction @xmath152 , and for @xmath153 @xmath154 a wick polynomial of the free field , an ads - invariant renormalization of the interacting field @xmath108 has been achieved , and that the boundary limit of @xmath108 exists with a suitable deformation @xmath103 of the power of @xmath38 as in . under these assumptions we shall prove : * proposition 2.1 : * _ if the boundary limit @xmath107 of @xmath108 exists , then it is a scale covariant field with scaling dimension d^w_l=_j=1^l \|a_j\|+ ^w_l . if @xmath138 contains no derivatives , then @xmath107 is a conformally covariant scalar field . _ this is , of course , a variant of the central result in @xcite , that the boundary limit of a scalar ads field , if it exists , automatically inherits unbroken conformal symmetry . the proof given there describes the cft as a `` theory la lscher mack '' ( * ? ? ? 3 ) on the cone @xmath155 , or a covering thereof . we want to include here a proof that refers directly to the cft on @xmath1-dimensional minkowski spacetime @xmath102 , which is ( a chart of ) the _ projective _ cone @xmath156 , or a covering thereof . ( @xmath157 is also known as the dirac manifold @xmath158 . ) _ proof of prop . 2.1 : _ let @xmath159 be the unitary representation of @xmath144 on the fock space of the free klein - gordon field @xmath2 on ads , which implements also the conformal transformation of the boundary generalized free field @xcite . for the subgroup corresponding to conformal scale transformations on the boundary , we have adu()(z , x ) u()(z , x)u()^ * = ( z , x ) [ ads - trans ] and hence adu()w(x)=^_j \|a_j\| w(x ) for @xmath57 . by means of and we conclude adu()r_n,1(l(x_1), ;w(x))= ^_j \|a_j\| r_n,1(l(x), ;w(x ) ) at non - coinciding points ( using here that the interaction @xmath82 contains no derivatives of @xmath2 ) . since we assume that an ads - invariant renormalization has been achieved , this identity is maintained in the extension to coinciding points . in terms of the interacting fields , this gives adu()w_l(x)=^_j \|a_j\| w_l(x ) in the algebraic adiabatic limit . with that and we obtain [ d = delta : proof ] & & adu ( ) w_l(x)=_z0z^-^w_l adu ( ) w_l(z , x ) = + & & = ^_j \|a_j\|+^w_l _ z0 ( z)^-^w_l w_l(x , z ) = ^_j \|a_j\|+^w_l w_l(x ) . this proves the first assertion of the proposition . we are now going to investigate whether the conclusion applies to arbitrary ads - transformations . let @xmath160 be an ads - transformation , @xmath161 the conformal transformation induced by @xmath162 on the boundary , i.e. , @xmath163 . for free wick powers @xmath138 ( without derivatives ) and consequently @xmath164 we obtain : [ wick : conf ] adu(\|t ) w(x)=_z0 z^-l adu(t)w(z , x ) = + = _ z0()^l ( z)^-l w(z,x ) = _ z0()^l w(\|t x ) , ( this argument would fail if @xmath60 involved derivatives . ) now , ads - invariance of the volume element @xmath165 implies z^-d-1 = z^-d-1 [ vol : invariant ] from which it is an easy exercise to conclude that in the limit @xmath166 ( where @xmath167 and @xmath168 ) one obtains _ z0 = ^1/d . thus , the factor in equals the conformal prefactor for a covariant field of scaling dimension @xmath31 . turning to interacting fields @xmath108 for @xmath138 , the ads - invariance of the retarded products , adu(t)r_n,1(l(x_1), ;w(x))= r_n,1(l(tx_1 ) , ; w(tx ) ) [ r : ads ] for @xmath169 , implies ads - invariance of the interacting bulk fields in the algebraic adiabatic limit the function @xmath170 does not factorize as if @xmath97 does ; but this does not obstruct our procedure thanks to prop . 8.1 in @xcite : in the algebraic adiabatic limit only the constancy of @xmath97 in the region of interest matters , and this is preserved by the transformation @xmath162 . ] : adu(t)w_l(x ) = w_l(tx ) . with that we find as before that @xmath101 is conformally covariant with scaling dimension @xmath171 ( provided it exists ) . namely , adu(\|t)w_l(x ) = _ z0 z^-_l^w adu(t ) w_l(z , x ) = + = _ z0()^^w_l(z)^-^w_l w_l ( z,x ) = ^_l^w / dw_l(\|t x ) . this completes the proof of prop . we want to illustrate by a simple model that renormalization of a field in @xmath3 dimensions and subsequent restriction to a @xmath1-dimensional submanifold is not equivalent to renormalization of the restricted fields . instead of @xmath158 as boundary of @xmath18 , we study the @xmath172-dimensional minkowski space @xmath173 ( with coordinates @xmath174 and relative coordinates @xmath175 ) as a submanifold of the @xmath176-dimensional minkowski space @xmath177 ( with coordinates @xmath178 and relative coordinates @xmath179 ) . the boundary limit corresponds to the restriction to @xmath173 of the fields in @xmath177 . the two - point function of a klein - gordon field of mass @xmath180 in @xmath102 is given by ^+(d)_m(y)(,(x+y)(x))=d^dp(p^0 ) ( p^2-m^2)e^-ipy [ delta^+ ] replacing @xmath181 and replacing @xmath175 by @xmath179 , this can be viewed as the 2-point function of a generalized free field in @xmath173 with @xmath182-dependent kllen - lehmann weight @xcite : ^+(5)_m(y ) = _ m^2^dm^2^+(4)_m(y ) . [ 5 - 4:2-point ] for later reference , we also introduce the corresponding commutator functions ^(d)_m(y)^+(d)_m(y)-^+(d)_m(-y ) [ d_m^d ] and the retarded propagators _ m^(d)(y)^(d)_m(y)(y^0 ) = d^dp , [ d_m^ret - d ] such that @xmath183 . we first investigate the renormalization of the fish diagram ( fig . 1 ) in @xmath177 . this means that we have to extend the distribution [ 5 - 4:r^0commu ] r_fish^(y)-i ( , [ , ] ) ( -y^0 ) = + = -2i(^+(5)_m(y)^2 - ^+(5)_m(-y)^2)(-y^0 ) , which is well defined for @xmath184 ( because @xmath185 $ ] vanishes for @xmath186 ) , to a distribution @xmath187 ( i.e. , to @xmath188 ) . the extension has to be such that it does not increase the scaling degree with respect to @xmath189 . to obtain a solution of the extension problem in @xmath177 , we work with the kllen - lehmann representation in @xmath177 . the square of the 2-point function is given in app . [ kallen ] . choosing for simplicity the field to be massless , this gives ( using with @xmath181 and @xmath190 ) r_fish^(y)=_0^dm^2 mi_m^(5)(-y ) . [ mink-5:r^0 ] the uv divergence of the unrenormalized distribution @xmath134 shows up in the divergence of the mass integral . the most general @xmath191 lorentz invariant extension with the required scaling degree is given by @xcite [ flatrenorm ] r_fish^()(y)(-_y+^2)dm^2 i_m^(5)(-y)(^20 ) depending on a renormalization parameter @xmath192 . ( the symbol @xmath193 stands for suppressed numerical factors ) . we have obtained @xmath194 by renormalizing in @xmath177 . we now consider how this distribution would appear when regarded as a distribution on the hypersurface @xmath173 with the transverse difference coordinate @xmath182 as a parameter . writing @xmath179 and the five - momentum as @xmath195 , we arrive at r_fish^()(u , y ) ( -_y+^2)dm^2 d^4p e^ipydv + ( -_y+_u^2+^2 ) d^4p e^ipydm^2 [ holo - flatrenorm ] the appearance of the derivative @xmath196 ( outside of the integrals ) is characteristic for the @xmath176-dimensional renormalization . one can not get rid of this operator , because it can not be shifted under the integral . ( the integrand is not differentiable with respect to @xmath182 at @xmath197 ) . it is the reason why @xmath176-dimensional renormalization `` as seen from the hypersurface '' goes beyond standard @xmath172-dimensional renormalization . one way to understand this fact is that on the hypersurface , the fields @xmath198 are independent fields which `` mix '' with @xmath199 upon 5-dimensional renormalization . in order to exhibit this more clearly , we compare the result of renormalization in the bulk with the alternative procedure of renormalization on the hypersurface , where we have a @xmath38-dependent family of fields in four dimensions , similar as in . the label @xmath38 just distinguishes different generalized free fields @xmath200 on the same hypersurface , see @xcite . that is , we write the @xmath176-dimensional 2-point functions in the unrenormalized distribution @xmath134 as a @xmath201-dependent integral over @xmath172-dimensional 2-point functions as in ( with @xmath202 ) , and apply the kllen - lehmann representation for the resulting products of 2-point functions as in with @xmath203 . this gives r_fish^(y)=-_0^dm_1 ^ 2_0^dm_2 ^ 2 [ 5 - 4:r^0 ] + ( ^+(4)_m_1(y)^+(4)_m_2(y)- ^+(4)_m_1(-y)^+(4)_m_2(-y))(-y^0 ) = + = _ 0^dm^2 f(m^2,u)i _ m^(4)(-y ) , with f(m^2,u)_0^dm_1_0^dm_2(m - m_1-m_2)[5 - 4:f ] + ( m_1u)(m_2u ) . the unrenormalized distribution @xmath134 exists in @xmath204 , but for @xmath205 , the mass integral on the right hand side of diverges in the region @xmath206 . renormalization on @xmath173 means regarding as a @xmath182-dependent kllen - lehmann representation in @xmath173 and extending it to the diagonal of @xmath173 in an @xmath207 lorentz invariant way . at @xmath208 , an extension to @xmath209 is in fact trivial because is already defined there , but the extension is non - unique ( @xmath210-functions in @xmath175 ) . in order to extend also to @xmath197 ( @xmath197 corresponds to two fields on the same hypersurface ) , one have to consider the most general @xmath207 lorentz invariant 4-dimensional renormalization [ submani ] r_fish^ ( ) ( y):= ( -_y+^2)dm^2 i_m^(4)(-y)(^20 ) . these distributions exist even in @xmath211 , have scaling degree @xmath212 and agree with @xmath134 for @xmath213 . so , @xmath214 solves the renormalization ( i.e. extension ) problem in @xmath173 . but it is not a renormalization in @xmath177 because it does not agree with @xmath134 at @xmath215 . to see this , we evaluate both and on a test function @xmath216 with @xmath217 . suppressing irrelevant constants , the difference is r_fish^()(g)-r_fish^(g ) dm^2 du ( u ) f(m^2,u)d^4k g(k ) ( - ) + [ difference ] dm^2 du ( u ) d^4k g(k ) g(0 ) du ( u ) f(m^2,u ) . one can actually compute @xmath218 by using variables @xmath219 and @xmath220 in , giving @xmath221 . thus , since @xmath217 , the @xmath182-integral in decays @xmath222 due to the oscillatory behaviour of @xmath223 , so that the @xmath224-integral is finite as required for a 4-dimensional renormalization . but it obviously does not vanish for generic @xmath93 , as would be required by a 5-dimensional renormalization . this proves the claim . note that the scale - invariant choice @xmath225 does not alter the conclusion ( the mass integral in in this case is @xmath226 ) . an analogous but more refined argument shows , that also when one admits a function @xmath227 , the resulting distribution can not coincide with @xmath134 for all @xmath228 . the fact that renormalization performed on a submanifold ( eq . ) does not coincide with proper renormalization in the bulk ( eqs . , ) , is the main message of this subsection . the breakdown of the bulk symmetry in the hypersurface renormalization is the counterpart of conformal symmetry breaking in ads - cft . it can be avoided by bulk renormalization , and subsequent restriction ( boundary limit ) . we proceed with some case studies concerning the compatibility of an ads - invariant renormalization with the existence of the boundary limit . we shall not endeavour the greatest possible generality ; e.g. , we shall always assume the ads mass parameter @xmath230 to be sufficiently large to avoid the breitenlohner - freedman critical behaviour in the range @xmath231 ( see , e.g. , @xcite ) . we start with the perturbative construction of the interacting field @xmath232 with interaction @xmath233 as a deformation of @xmath0 . the renormalization of @xmath234 in this case is unproblematic , but it serves to illustrate the difference between various approaches . in order to work out the boundary limit of the renormalized bulk field @xmath235 , we introduce a general technique of computation ( sect . [ sec : massshift ] to be used in more general cases as well . in the subsequent section , we shall choose to study the renormalization and boundary limit of the field @xmath236 because in this case , the perturbative expansion involves a loop diagram ( the fish diagram , fig . 1 ) already at first order . our strategy is to construct the interacting ads field @xmath237 , and then take its boundary limit . in the diagrammatic expansion of @xmath237 , each diagram has a single propagator line extending from @xmath238 to the first interaction vertex @xmath239 ( fig . 2 ) . therefore , the @xmath17 behaviour of each diagram is dictated by the same function ( apart from potential ir problems ) , so that the analysis of the limit can be essentially done in the first order . nontrivial renormalization , in contrast , becomes relevant only at higher order . * fig . 2 : * factorization of @xmath240 . to first order perturbation theory @xmath241 we obtain _ ^k^(1)(x)=k _ 0^ ( z_1 ) d^dx_1 g(x_1)i_ads^ret(x , x_1 ) , [ phi_l^1 ] where @xmath242 is the retarded propagator on @xmath3-dimensional ads , according to given by ^ret_ads(x , x_1 ) = 12(zz_1)^d/2 dm^2 j_(mz)j_(mz_1)^ret(d)_m(x - x_1 ) . [ delta^ret_ads ] at this point , one might be tempted to read off the @xmath166 behaviour directly from and the well - known behaviour of the bessel functions near zero . we shall see , however , that this attempt is too naive , and that the subsequent @xmath243-integration in changes the limit behaviour substantially . for the trivial case @xmath245 ( in which the `` interaction '' amounts just to a shift of the field by a constant ) , the adiabatic limit @xmath246 , @xmath247 can be taken directly in and yields the expected result ^(1)_(x ) = d^dx_1i^ret_ads(x , x_1 ) = 1m^2 , [ phi^1:k=1 ] which follows from @xmath248 upon integration over @xmath239 , using ads - invariance so that the integral does not depend on @xmath238 . one may also perform the integrations explicitly in the representation where the @xmath249-integration is obvious from , and the subsequent @xmath243- and @xmath250-integrations are carried out using formula ( 13.24(1 ) ) in @xcite , _ 0^duu^j_(u ) = 2^ ( --1 < z ) + o(^2 ) , [ log : anodim ] where the first order term is [ phi_1:k=2 ] ^(1)_^2(x ) = 2(z_1)d^dx_1g(x_1)i^ret_ads(x , x_1)(x_1 ) . indeed , in the partial adiabatic limit @xmath251 exhibits a logarithmic @xmath38-dependence which is precisely cancelled by the combination occurring in . namely , implies ( _ x+m^2 ) ^(1)_^2(x ) = 2(z)g(x)(x ) and consequently , using ( _ x+m^2 ) ( ^(1)_^2(x)+ ( x ) ) = 2(-z_z)(x ) [ inhomogen ] in the region where @xmath252 , @xmath253 . the right - hand side vanishes in the limit @xmath166 faster than @xmath254 because the leading @xmath254 behaviour of the unperturbed field is annihilated by the differential operator @xmath255 . since the klein - gordon operator preserves homogeneity in @xmath38 ( except for the @xmath256 term which is suppressed at small @xmath38 ) , the combination of fields on the left - hand side also vanishes faster than @xmath254 , up to a solution of the homogeneous equation . the homogeneous solution can behave @xmath257 or @xmath258 . if we can exclude the latter ( dominant ) contribution , then it follows that the limit at first order in @xmath13 exists . unfortunately , the previous argument based on the klein - gordon operator can not discriminate between @xmath257 and @xmath259 . we shall therefore develop a more refined analytical method of computation which is `` universal '' ( see lemma b.1 in app . [ app : lemma ] ) in the sense that it can also be applied when dealing with interactions of higher polynomial degree ( sect . [ sec : polynomial ] ) and with diagrams with loops ( sect . [ sec : phi2 ] ) . this method at the same time shows the emergence of the @xmath260 terms . the argument is lengthy , with essential parts contained in app . [ app : lemma ] , but it is crucial for the understanding of the boundary limit . for the sake of transparency and computational simplicity , we present only the case d=3 the ads 2-point functions are explicitly known in terms of hypergeometric functions or associated legendre functions of the second kind @xcite : let @xmath261 , @xmath262 ( @xmath263 ; @xmath264 ) , and v= . @xmath265 is ads - invariant . namely , viewing @xmath18 as the hypersurface @xmath20 in a @xmath266-dimensional ambient space @xmath22 of signature @xmath21 , we have v=_1 , hence @xmath265 is related to the `` chordal distance '' by @xmath267 . we expect singularities at @xmath268 ( @xmath269 ) and , due to the the identification of @xmath270 with @xmath271 , also at @xmath272 ( @xmath273 ) . note also that timelike separation between @xmath238 and @xmath239 corresponds to @xmath274 $ ] . then for @xmath275 [ delta_ads ] ^+_ads(x_1,x ) = - q_- 12(v+iy^00 ) . here @xmath276 is a solution of legendre s differential equation ( 1-u^2)f - 2u f + ( + 1)f=0 , which is analytic outside a cut along the real interval @xmath277 $ ] . for @xmath202 , hence @xmath278 , @xmath279 , it is the elementary function [ q1 ] q_1(u)= u2 - 1 q_1(u)=12(1+u_u ) . the retarded propagator @xmath280 is given by the discontinuity across the cut : 4i^ret_ads(x , x_1)&= & 12(1+u_u)^u = v+i0_u = v - i0(-y^0 ) + & = & -(1+v_v ) ( 1-v)(-y^0 ) . this discontinuity is to be understood as a distribution by partial integration w.r.t . @xmath265 : [ hdef ] h[f]:= -dv f(v)(1+v_v)(1-v ) = _ -1^+1 dv v_v f(v ) . because we have represented the retarded propagator as a distribution w.r.t . the variable @xmath265 , we have to perform all other integrations ( at fixed value of @xmath265 ) first . we therefore change the integration variables : in spatial polar coordinates , let @xmath281 , and @xmath282 . then the new variables are v , z_1,t -y^0 , . the measure becomes d^3y(-y^0 ) ( z_1 ) = zdv ( z_1 ) dt(t)(t^2-w)d , where [ wz ] w = w_v , z(z_1)= z^2+z_1 ^ 2 - 2vzz_1 ( z_1v - z)^2 + ( 1-v^2)z_1 ^ 2 . there is a dense domain of vectors for which matrix elements @xmath283 of the distributional field become smooth function . we then extract the leading @xmath243 behaviour and write ( z_1,x^0-t , x+re_):=(z_1)g(x_1)z_1 ^ -3 ( _ 1,(z_1,x_1)_2 ) . this is a smooth function with compact support , because of the cutoff functions @xmath92 and @xmath93 . at @xmath284 , it equals the corresponding matrix element of @xmath28 , because @xmath247 and @xmath246 in the region of interest ( partial adiabatic limit ) . finally we average over the spatial directions and put _ x(z_1,t , r^2):= 12d ( z_1,x^0-t , x+re _ ) . then @xmath285 is smooth ) that @xmath285 is regular in the _ quadratic _ variable @xmath286 . this is obvious at @xmath287 because the square root is smooth . at @xmath288 , the smoothness can be seen by a taylor expansion with remainder of @xmath289 , because the angular averaging annihilates all odd terms . ] in all three arguments @xmath290 , and [ gx0 ] _ x(0,0,0 ) = ( _ 1,(x)_2 ) . with these preparations , ( the matrix element of ) the first - order correction to the renormalized field becomes [ hgamma ] ( _ 1,^(1)_^2(x)_2 ) = zh , with the functional @xmath291 $ ] as defined in . we claim , that this equals ( _ 1,^(1)_^2(x),_2 ) = -23 z^3(z_x(0,0,0 ) + ) , [ claim ] where ( regular ) stands for a contribution that is regular in @xmath38 at @xmath85 . the argument goes as follows . for a smooth function @xmath292 on @xmath293 with compact support , we denote by @xmath294 the integral [ i0vz ] i_0(v , z)(f):= _ 0^dz_1 _ 0^dt(t^2-w ) f(z_1,t , t^2-w)_w = w_v , z(z_1 ) . thus , to compute , we have to apply the functional @xmath75 to @xmath294 when @xmath292 equals @xmath285 on @xmath295 . in app . [ app : lemma ] , we prove that @xmath294 is continuous w.r.t . @xmath265 and differentiable in the range @xmath296 . thus , the definition of @xmath75 by partial integration is unambiguous , and it is sufficient to know this function at @xmath296 , where @xmath297 . in physical terms , this remark means that there are no singular contributions from lightlike @xmath175 ( @xmath298 ) : the integration can be properly computed by exhausting the backward lightcone `` from the inside '' . in app . [ app : lemma ] , we also prove that in the range @xmath296 , @xmath294 is of the form i_0(v , z)(f)= -2mm_0k2 -3 mm a_k(f)v^kz^+ z^22 ( ( 1-v)z)f(0,0,0 ) + r_v , z(f ) where @xmath299 are certain distributions that do not depend on @xmath265 and @xmath38 , while the remainder @xmath300 is a family of distributions that is differentiable w.r.t . @xmath265 in the range @xmath296 , and vanishes @xmath301 at @xmath85 . noting that @xmath302=h[v^1]=0 $ ] , the leading terms are annihilated : h & = & -2z^2z f(0,0,0)+ + & + & z^2(a_22(f ) + hf(0,0,0 ) ) + h[r_v , z(f ) ] . thus , with @xmath303=\frac 43 $ ] , we have * proposition 3.1 : * _ for any test function @xmath292 on @xmath293 , the limit _ z0 z^-2 \{h + 23z^2z f(0,0,0 ) } is finite . _ for @xmath304 on @xmath295 , this is our claim . this ensures that @xmath305 decays at least like @xmath306 , and because of , it also ensures that @xmath307 ( recall @xmath278 in ) decays at least like @xmath308 . in other words , the boundary limit exists ( in first order perturbation theory , and in the obvious weak sense ) , and is exactly given by the expected correction of the scaling dimension of the boundary field . apart from establishing the existence of the ( expected ) boundary limit , the main message to be drawn from the computation in app . [ app : lemma ] , however , is that * _ the origin of the logarithmic term ( corresponding to the anomalous dimension ) is the range @xmath309 of the integral , and not the power law behaviour of the retarded propagator at @xmath85 . _ we now turn to the non - trivial interactions @xmath310 . in these cases yields ( _ x+m^2 ) ^(1)_^k(x ) = ( z)g(x)k where the right - hand side @xmath311 vanishes faster than @xmath254 . by the same argument as used after , @xmath312 behaves either like @xmath254 or like @xmath313 . in the special case @xmath275 , @xmath202 , we can explicitly see the absence of the `` wrong '' contribution @xmath258 , by repeating the explicit computation as in the previous section . replacing @xmath314 by @xmath315 , one gets an additional factor @xmath316 in . because the logarithmic term in this case appears at order @xmath317 ( lemma b.1 ) , it is manifest that the first - order term is of order @xmath318 , as desired , and there is no logarithmic term . thus , the boundary limit exists without an anomalous dimension . although a complete analysis of renormalization at higher - order is beyond the scope of this paper , let us anticipate what happens in the case at hand . first , we observe ( see fig . 2 above ) that @xmath319 can be written as _ ^k^(n ) = k _ 0^ ( z_1 ) d^dx_1 g(x_1)i ^ret_ads(x , x_1 ) ( ^k-1)_^k^(n-1)(x_1 ) . thus , in order to renormalize @xmath320 at order @xmath126 , one previously has to renormalize @xmath321 at order @xmath322 . in principle , one has to renormalize `` all fields simultaneously '' , but in practice , for any finite order of any given field it is sufficient to renormalize only a finite number of fields to lower orders . thus , assuming recursively that @xmath323 has been defined up to @xmath324st order , and anticipating that its boundary limit exists with an anomalous dimension of order @xmath150 , then @xmath325 behaves like @xmath326 times a polynomial in @xmath327 , as @xmath17 . because the canonical dimension @xmath328 is larger than @xmath14 , the same argument as before applies to ensure that the partial adiabatic limit for @xmath320 is unproblematic , and for @xmath38 sufficiently small ( such that @xmath252 ) , the equation ( _ x+m^2 ) ^(n)_^k(x ) = kg(x ) ( ^k-1)_^k^(n-1)(x ) implies the @xmath254 behaviour of @xmath329 as @xmath17 . again , this equation does not yet exclude a term @xmath258 , but an explicit computation as in lemma b.1 in the special case @xmath275 , @xmath202 again shows its absence . we conclude that anomalous dimensions do not arise also in higher orders of perturbation theory . actually , one can go beyond this statement : even if the logarithms could be summed ( borrowing suitable higher order terms , i.e. , violating the proper perturbative systematics ) to give rise to an anomalous dimension @xmath330 up to order @xmath322 ( see sect . [ sec : phi2 ] ) , then the argument would still hold true as long as @xmath331 ( cf . lemma b.1 with @xmath332 ) . in the next section , we shall discuss the behaviour of `` composite fields '' @xmath333 . depending on the interaction , these fields will exhibit finite anomalous dimensions . we conclude this section with a comparison of the competing renormalization prescriptions in the case at hand . concerning the renormalization , we find here significant differences between ( a ) our procedure , as just outlined , and ( b ) perturbation theory around the generalized free field @xmath0 in minkowski space @xmath102 , requiring poincar invariance ( b1 ) , or in conformal minkowski space @xmath158 , requiring conformal invariance ( b2 ) : * ( renormalization in the bulk ) the numerical distribution @xmath334 @xmath335 coincides with the retarded propagator @xmath336 in @xmath23 . its extension to the diagonal is uniquely given by , and there is no freedom of renormalization , because its scaling degree in the relative coordinates equals @xmath337 ( for @xmath338 ) , which is smaller than the dimension of the relative coordinates ( @xmath339 ) @xcite . the boundary behaviour of the resulting fields is dominated by the @xmath243-integration near @xmath309 , which depends sensitively on the operator valued distribution with which @xmath340 is multiplied . it is important to keep in mind that we have _ renormalized ( extended @xmath341 to the diagonal ) first , and then taken the limit @xmath17 ( in the partial adiabatic limit at the boundary)_. * ( renormalization on the boundary ) doing perturbation theory on the boundary , instead , we have to take the limit @xmath17 first . this yields the unrenormalized distribution @xmath342 : r_^(x - x_1 ) = i[(x),(x_1)](x^0-x_1 ^ 0 ) = dm^2 m^2 i_m^(x - x_1 ) . [ retprop : m ] this product of distributions exists on @xmath343 only in the range @xmath344 . . alternatively , it can be obtained by taking the boundary limit @xmath345 of . this limit may be done before the mass integration in iff @xmath344 . ] for @xmath346 the integral @xmath347 diverges , nevertheless @xmath348\theta(x^0-x_1 ^ 0)$ ] is well defined for @xmath349 , and one is faced with the problem to extend @xmath350 from @xmath351 to @xmath343 . one has two options : * * case ( b1 ) : one only requires that the lorentz invariant extension does not increase the scaling degree ( with respect to @xmath352 ) of @xmath350 @xcite , which has the value @xmath353 . in this case , the retarded propagator is _ non - unique _ for @xmath346 : the general solution reads r_(y)=(^2-_y)^[]+1dm^2+_n c_n^n(y ) [ retprop : m - int ] where @xmath354 and the @xmath355 s are arbitrary constants ( cf . appendix c of @xcite ) . clearly , the renormalization mass @xmath192 and the local terms break the scale invariance ( unless @xmath356 ) . * * case ( b2 ) : requiring conformal covariance of the extension , a necessary condition is that the homogeneous scaling behaviour of @xmath357 is maintained : this is an intensification of the requirement in ( b1 ) . from we see that there is a unique solution for @xmath358 which is obtained by choosing @xmath359 and @xmath360 . but if @xmath361 , the mass integral is ir - divergent for @xmath359 , and a scaling covariant retarded propagator does not exist . we turn to the field @xmath236 with interaction @xmath116 ( @xmath363 ) . in this case , there exist three types of diagrams which a priori behave differently as @xmath17 : those diagrams in which the two interaction vertices connected to the field vertex are distinct and do not belong to a common loop , those in which they are distinct and belong to a common loop , and those in which they coincide . arising in perturbation theory for the interacting field @xmath132 with interaction @xmath116 . diagrams of the first type factorize into two diagrams as for the field @xmath235 and consequently can be treated as in sect . [ sec : polynomial ] . the second type does not arise in first order . diagrams of the last type contain the fish diagram ( fig . 1 ) as a subdiagram , which determines their @xmath38-dependence . this diagram gives the contribution to @xmath364 _ 0^(z_1)dx_1 g(x_1 ) r_fish(x_1;x ) . [ fish : intf ] in order to define this contribution , the unrenormalized distribution @xmath365 , given by r_fish^(x_1;x)-2i(^+_ads(x_1,x)^2 - ^+_ads(x , x_1)^2)(x^0-x_1 ^ 0 ) [ fish : nonren ] at @xmath366 ( cf . ) , has to be extended to the diagonal @xmath367 . then we have to study the boundary behaviour @xmath166 of the renormalized integral in the partial adiabatic limit . our task is to understand the influence of the uv renormalization on the boundary limit . the unrenormalized distribution is real - valued and ads - invariant . we require that the extension @xmath368 has the same properties : * @xmath369 is real - valued ( i.e. , @xmath370 ) and the scaling degree in the relative coordinates @xmath371 is not increased by the extension : _ y(r_fish ( ; x))=_y ( r_fish^ ( ; x))=2d-2x . [ fish : sd ] * @xmath369 is ads - invariant r_fish(tx_1;tx)=r_fish(x_1;x)tso(2,d ) . [ ads - inv ] in addition , we want to impose the existence of the boundary limit of the interacting field @xmath372 as a condition on the renormalization , admitting for an anomalous dimension @xmath373 . thus , up to first order of perturbation theory , + ( -z ) [ log : cancel ] should converge with @xmath166 . we have already seen that the contributions from the first type of diagrams ( fig . 3 ) to @xmath374 behave @xmath375 if @xmath310 , and with a logarithmic correction if @xmath376 , so that their limit exists separately because @xmath377 . because the only possibly divergent contribution comes from the fish diagram integrated with @xmath378 , a cancellation against the contribution from an anomalous dimension can occur in only if @xmath379 and only if the divergence of @xmath380 integrated with @xmath381 is logarithmic . thus , we are led to require * the renormalized expression taken in the partial adiabatic limit and multiplied by @xmath382 converges at @xmath166 if @xmath383 , while for @xmath379 it may diverge @xmath384 . due to general theorems @xcite there exist extensions which fulfill ( i ) and ( ii ) . for @xmath385 , these two requirements reduce the freedom of normalization to r_fish(x_1;x)+cz^d+1(x_1-x)(z_1-z ) . [ freiheit ] so there is only one normalization constant @xmath386 at disposal to fulfil ( iii ) . for this reason , we concentrate on @xmath387 and @xmath203 from now on . changing the value of @xmath386 just adds a multiple of @xmath378 to @xmath374 . if @xmath376 or @xmath388 , this term @xmath389 or @xmath257 must not be present in the boundary limit taken with @xmath382 , so condition ( iii ) if it can be fulfilled fixes the value of @xmath386 , and thus determines a `` field mixing '' . if @xmath379 , the addition just amounts to a multiplicative renormalization of the zero order term . if @xmath390 , the addition is ineffective in the boundary limit . in both cases @xmath391 , the renormalization parameter @xmath386 is unconstrained by condition ( iii ) . these a priori conclusions are in perfect agreement with the corresponding conclusions drawn from the analysis of witten diagrams for correlation functions in the dual approach to the ads - cft correspondence @xcite . the standard strategy @xcite to renormalize ( extend ) a distribution like the fish diagram @xmath134 in curved space - time is to pass to the scaling limit @xmath392 which gives a distribution in the tangent space at the point @xmath238 . the latter carries the leading uv singularity and can be renormalized as in flat space ( with the constant metric @xmath393 ) , while the less singular `` reduced '' distribution @xmath394 is ( in @xmath275 or @xmath203 ) uniquely extended `` by continuity '' . the problem with this strategy in our situation is that @xmath395 and @xmath396 ( the latter being independent of @xmath14 because the scaling limit looses the information about the ads mass @xmath230 ) behave differently at the boundary , and do not allow us to deduce the boundary behaviour of the integral . let us look more closely at the distribution . unfortunately , the ads kllen - lehmann expansion of @xmath397 is not known explicitly @xcite , with which one could perform the renormalization in the spirit of . instead , we shall use again the explicit form of @xmath398 in @xmath399 bulk dimensions , and its elementary expression if @xmath202 , hence @xmath278 and @xmath279 . in order to renormalize ( i.e. , to define the retarded product as a distribution on @xmath400 ) , we adopt the method of differential renormalization @xcite : as a distribution on @xmath401 , is of the form r^_fish(x_1,x ) = j(x_1,x)(x^0-x_1 ^ 0 ) with @xmath402 . one writes j(x_1,x)=_x_1j(x_1,x ) where @xmath403 is an @xmath23-invariant distribution which vanishes if @xmath239 is spacelike separated from @xmath238 , and @xmath404 , so that @xmath405 is well - defined as a distribution on @xmath406 . one then defines r_fish(x_1,x ) : = _ x_1(j(x_1,x)(x^0-x_1 ^ 0 ) ) . at @xmath407 , this differs from the unrenormalized distribution @xmath408 by a term _ 0(j(x_1,x)(x^0-x_1 ^ 0 ) ) + ( x^0-x_1 ^ 0 ) _ 0 j(x_1,x ) . the support property of @xmath403 ensures that this vanishes at @xmath407 , hence @xmath409 is indeed an extension of @xmath410 . obviously , @xmath369 satisfies the requirements ( i ) and ( ii ) above . we follow this strategy in the case @xmath202 , where by , @xmath411 is given explicitly in terms of the elementary function . we thus obtain ( ^+_ads(x_1,x))^2 = + = ( ( ) ^2 + u_u ( ) ^2 + ( - ) ^2)_u = v+iy^00 . here , the first term is a logarithmically bounded function , hence well - defined as a distribution , and consequently also the second . the last term is defined as a distribution by ( 1v1+iy^00)^2 = -_v ( 1v1+iy^00 ) . we now look for a _ function _ @xmath412 such that [ c1 ] _ x_1 f(u ) ( 1-u^2)f(u)-4u f(u ) = q_1(u)^2 and [ c2 ] j(x_1,x)=(f(v - iy^00)-f(v+iy^00))=0v>1 . next , we determine the discontinuity along the cut f(v ) = f(v+i0 ) - f(v - i0 ) as a _ distribution_. then , we can define the renormalized fish diagram as [ fish : ren ] r_fish(x_1,x)=_x_1(f(v)(-y^0 ) ) . * proposition 4.1 : * _ equation is solved by [ f(u ) ] f(u)&=&12(_3+_3)+ 16(_3-_3 ) + & & + 16 ( _ 2 -_2)-116 ( ) ^2 + & & + 1144(u^2 + 3)()^3 + ( ) ^2 + plus the general solution @xmath413 of the homogeneous equation . _ we point out that @xmath414 is analytic for @xmath415 $ ] ( see appendix [ app - pf ] ) , and that the particular solution given by the proposition is symmetric ( @xmath416 ) , but @xmath417 . by writing some terms as derivatives , the boundary values @xmath418 are defined as distributions . _ proof : _ by insertion into . in appendix [ app - pf ] we sketch the derivation of . * proposition 4.2 : * _ the discontinuity @xmath419 is given by [ df ] f(v ) = i((1-v)h_0(v ) + _ v ( ( 1-v)h_1(v ) ) ) , where _ h_0(v ) & = & 13(_22 - _ 22 ) + 23 , + h_1(v)&= & 13 22 -v4 + ( + 512 ) . notice that the derivative of @xmath420 can not be taken separately , because @xmath421 is logarithmically divergent at @xmath422 . instead , @xmath423 is understood as a distribution in @xmath265 , where the derivative is defined by partial integration , see below . _ proof : _ see appendix [ app - pf ] . adding the homogeneous solutions , the second of the integration constants , @xmath424 , does not contribute to the discontinuity . thus , the ( expected ) renormalization freedom consists in adding to the term [ renfree ] & & c_1_x_1((q_1(v - iy^00)-q_1(v+iy^00 ) ) ( -y^0 ) ) = + & & = c_1_x_1^ret_ads(x , x_1 ) = -z^4(z_1-z)^(3)(x_1-x ) . * remarks : * ( i ) in contrast to the renormalization of the massless fish diagram in 4-dimensional minkowski space , the present _ renormalization on ads does not require the introduction of a mass scale . _ this is because there is already a mass scale in the formalism , namely @xmath425 , where @xmath426 is the radius of ads . ( in our conventions : @xmath427 . ) + ( ii ) @xmath423 in prop . 4.1 is _ antisymmetric _ in @xmath265 ; however the renormalization freedom is _ symmetric _ in @xmath265 . hence , there is a _ distinguished renormalization _ : @xmath428 . the term contributes a multiple of @xmath429 to the first order term of @xmath430 . as discussed in sect . [ sec : general ] , for @xmath390 this terms does not contribute to the boundary limit , while for @xmath379 its boundary limit @xmath431 exists trivially and amounts to a multiplicative renormalization of @xmath432 . for @xmath388 , it produces a `` mixing '' of the field @xmath433 with @xmath434 , and the boundary limit has to be taken of the appropriate mixed field ( cf . the end of the sect . [ sec:3limit ] ) . we shall therefore disregard this term in the sequel . thus , with @xmath423 specified by prop . 4.2 is the starting point for the subsequent analysis of the boundary limit . in that analysis , @xmath423 is understood as a distribution on the differentiable functions on the interval @xmath435 , i.e. , [ hfish ] h_fish[f ] 1i f[f ] : = _ -1^+1 dv f(v ) . the crucial property will be * proposition 4.3 : * _ the linear functional @xmath436 vanishes on even powers @xmath437 , and h_fish(v^2m+1 ) = 23 _ = 0 ^ 2m+1 j_+ 6 j_2m+1 - 56 where @xmath438 are given in . in particular , @xmath439=0 $ ] for @xmath440 , and @xmath441 = \frac 4{81}$ ] . _ _ proof : _ the even powers of @xmath265 are automatically annihilated by @xmath436 by symmetry under @xmath442 . for the odd powers , see appendix [ app : limit ] . let us first consider the most interesting case of the interaction @xmath443 , i.e. , @xmath379 . the fish diagram contribution to the first order correction to @xmath430 is given by [ firstorder ] & & 6 d^3x_1 g(x_1)_0^ ( z_1)r_fish(x_1,x ) = + & & = d^3y _ 0^ f(v)(-y^0 ) _ x_1((z_1 ) g(x+y ) ) , where @xmath444 , @xmath445 , and @xmath446 as before . to study the boundary limit , we proceed exactly as in sect . [ sec : massshift ] , when evaluating . we choose again @xmath275 and @xmath202 . making the same change of variables , we put ( z_1,x^0-t , x+re_):= z_1 ^ -6 _ x_1((z_1 ) g(x_1 ) ( _ 1,_2 ) ) and _ x(z_1,t , r^2):=12d(z_1,x^0-t , x+re _ ) . again , @xmath285 is regular at @xmath352 , and [ g0 ] _ x(0,0,0)=-18(_1,_2 ) . the factor @xmath447 is produced by the laplace operator when acting on @xmath448 at small @xmath243 . then we arrive at the matrix element of [ firstorder2 ] = -h_fish which is of the same form as , except for the additional power @xmath449 ( due to the factor @xmath433 in as compared to @xmath2 in ) , and with the functional @xmath75 replaced by @xmath436 given in . the argument in square brackets is of the form @xmath450 with @xmath304 on @xmath295 , as computed in the lemma b.1 of app . [ app : lemma ] . by the same arguments as before , it is sufficient to know it in the range @xmath296 , where it is given by : there are polynomial terms @xmath451 , a logarithmic contribution [ logterms ] z^5b_3(v ) ( ( 1-v)z ) f(0,0,0 ) b_3(v)= 8v(1-v^2)(7v^2 - 3 ) , and a remainder @xmath452 that vanishes in the boundary limit @xmath17 . by prop . 4.3 , the leading polynomial terms with @xmath453 are annihilated by @xmath454 $ ] , so that only the term @xmath441\,a_{55}(f)\,z^5 $ ] survives . the @xmath455 term in produces another constant in @xmath456 in lemma b.1 ensures the finiteness of @xmath457 $ ] . ] times @xmath458 , and the @xmath327-term produces the contribution @xmath459 \cdot z^5\log z \cdot f(0,0,0)$ ] . since @xmath441=\frac 4{81}$ ] ( prop . 4.3 ) , we have thus found the following analog of prop . 3.1 : * proposition 4.4 : * _ for any test function @xmath292 on @xmath293 , the limit _ z0 z^-5 \{h_fish+ 7162z^5z f(0,0,0 ) } is finite . _ inserting this result with @xmath304 and into , we find the first order contribution = -(z + o(z ) ) . the absence of all lower order terms establishes the existence of the boundary limit , and the presence of the logarithmic term signals the anomalous dimension of the composite boundary field ^^2_^2 = 6 - 76 ^ 2 + o(^2 ) . at first order of perturbation theory . _ this establishes the existence of the boundary limit of @xmath432 in first order perturbation theory , when @xmath202 . _ the result requires the nontrivial cancellations @xmath460=h_{{\rm fish}}[v^3]=0 $ ] of prop . 4.3 , involving the precise functions @xmath461 and @xmath421 of prop . 4.2 appearing in the renormalized fish diagram . it remains to investigate whether similar cancellations persist for @xmath462 , @xmath463 , and at higher orders . it is now easy to repeat the analysis for the interaction @xmath464 , i.e. , @xmath465 on the r.h.s . of has to be replaced by @xmath466 . in this case , the power @xmath449 in the @xmath243-integral is absent ( @xmath467 in lemma b.1 ) , hence the logarithmic term @xmath327 arises at order @xmath468 with a coefficient @xmath469 . because @xmath436 annihilates the quadratic polynomial @xmath470 , but not @xmath471 , the first - order diagram will not contain @xmath327 terms , but finite terms @xmath472 . this reflects the expected perturbative mixing of the fields @xmath473 and @xmath2 under the cubic interaction . accordingly , the boundary limit @xmath17 should be taken of a suitable combination like @xmath474 . we have pursued the strategy of perturbative construction of interacting conformal fields in @xmath1 dimensions , which proceeds by the perturbative construction of interacting ads fields in @xmath3 dimensions and subsequently performing a boundary limit . the unperturbed conformal field is a generalized free field ( or a wick product thereof ) . this procedure resolves the problematic issues associated with the perturbation theory around generalized free fields , and at the same time drastically reduces the expected infinite arbitrariness involved in its renormalization . the most important benefit is the fact , that the boundary fields , if renormalized by this method , do not suffer from the conformal anomaly , i.e. , the conformal symmetry is perturbatively preserved . we find , however , that the existence of the boundary limit is not automatically guaranteed . requiring its existence may be viewed as another renormalization condition for the ads field which can not always be fulfilled . we have pursued a number of case studies involving polynomial interactions of scalar fields . in relevant cases , the boundary limit exists , and the renormalized boundary fields have anomalous dimensions that can be computed . ( an anomalous dimension does not mean a conformal anomaly ! ) because the exact analytical expressions are quite involved , we have considered only very special cases ; but in view of the highly systematic emergence of the cancellations , we believe that the promising results found in these cases pertain also to more general cases . the method is applicable only when the lagrangean interaction density of the conformal boundary field is induced by a polynomial interaction on ads . such densities are rather special elements of the borchers class of the generalized free field , which carry a reminiscence of its ads origin . but in view of the fact that a general perturbation theory for generalized free fields has not yet been formulated , it is encouraging that a successful renormalization can be achieved at least for a limited class of interactions . there arises an interesting question , concerning the `` continuous operator product expansion '' for generalized free fields , as discussed in sect . [ sec : renorm ] . the ope in the bulk is certainly a discrete sum . taking the boundary limit , when it exists , should not alter this feature . recalling that the continuous ope is caused by the failure of factorization of the weight functions @xmath475 in , we are tempted to conjecture the perturbative stability of a discrete ope for `` factorizing '' wick products whenever only the lagrangean is a non - factorizing generalized wick product . to establish such a result , one would have to reorganize the ope of the perturbed limit fields , whose subleading terms are continuous in terms of the unperturbed fields , into a discrete ope in terms of the perturbed fields . let @xmath477 denote the 2-point function of a massive scalar free field in @xmath1-dimensional minkowski space . we are going to prove ^+_m_1(y)^+_m_2(y)=_0^ dm^2_m_1,m_2(m^2 ) ^+_m(y ) [ kl ] with [ rho ] _ m_1,m_2(m^2 ) = & & + & & -5 mm ( m - m_1-m_2)m^2-d((m^2-m_1 ^ 2-m_2 ^ 2)^2- 4m_1 ^ 2m_2 ^ 2)^ where @xmath478 is the surface of the unit sphere @xmath479 in @xmath3 dimensions , \|s^d\|= . from the definitions , and using lorentz invariance , it is easily seen , that the kllen - lehmann weight is given by _ m_1,m_2(m^2)&= & + & & -7mm_v_+d^dp_1 _ v_+d^dp_2(p_1 ^ 2-m_1 ^ 2 ) ( p_2 ^ 2-m_2 ^ 2)(p_1+p_2-p ) where @xmath480 is any four - momentum such that @xmath481 . it is convenient to choose @xmath482 and perform the integrations over the energies @xmath483 first , and evaluate the momentum conservation @xmath484 . the resulting integral over @xmath485 reads in polar coordinates @xmath486 _ m_1,m_2(m^2)&= & + & & -7 mm _ 0^ ( + -m ) . [ rho1 ] the argument of the @xmath210-function vanishes at p_0= provided @xmath487 and @xmath488 , where the first bound is redundant . from this , we obtain . we use notations as introduced in sect . [ sec : massshift ] , with @xmath489 . for a test function @xmath292 on @xmath293 , we denote by @xmath490 the integral [ iuvz ] i(u , v , z):= i(w)_w = w_v , z(u),i(w):=_0^dt ( t^2-w ) f(u , t , t^2-w ) , and by @xmath491 the integral [ invz ] i_n(v , z)(f):= _ 0^u^ndu i(u , v , z)(n0 ) . we want to prove : * lemma b.1 : * _ let @xmath492 . then @xmath491 is continuous w.r.t . @xmath265 . in the range @xmath296 , it is of the form [ udvz ] i_n(v , z)(f ) & = & _ 0kn+2 a_k(f)v^kz^+ + & + & b_n(v ) z^n+2 ( ( 1-v)z)f(0,0,0 ) + r_v , z(f ) , if @xmath493 is an integer . here , @xmath299 are distributions , @xmath494 is a polynomial of degree @xmath495 , and the remainder @xmath300 is a family of distributions that is differentiable in @xmath265 in the range @xmath296 , and that vanishes at least @xmath496 as @xmath166 . if @xmath497+\varepsilon$ ] is not an integer , then the first ( polynomial ) sum extends until @xmath498 + 2 $ ] , the logarithmic term is replaced by @xmath499 + 2+\varepsilon}\cdot f(0,0,0)$ ] with a possibly non - polynomial function @xmath355 , and the remainder is @xmath500 + 3})$ ] . _ * remark : * the emphasis is here on the various subleading terms after the polynomial terms , because they become the leading ones in different instances of our case studies of the boundary limit , and we expect that this happens also in more general cases . the @xmath327-term is essential for prop . 3.1 and prop . 4.4 . the @xmath455-term is used in the last paragraph of sect . [ sec:3limit ] , and the @xmath501 + 2+\varepsilon}$]-term in the non - integer case is relevant in sect . [ sec : polynomial ] . _ proof : _ the integrals @xmath490 and @xmath491 are continuous w.r.t . @xmath265 by definition , because the integrand and the range of integration vary continuously . for the differentiability w.r.t . @xmath265 when @xmath296 , we note that the dependence on @xmath265 is only through @xmath99 , and @xmath502 . thus @xmath503 , and [ dwiw ] -_w i(w ) = 12w f(u , w,0 ) + _ w^dt _ 3f(u , t , t^2-w ) . we now compute the leading derivatives w.r.t . @xmath38 in the range @xmath296 . again , the dependence is only through @xmath504 , and @xmath505 , hence [ dnuiw ] _ z^i(u , v , z ) = _ k=[2]^ c^_k(uv - z)^2k-(-_w)^ki(w)_w = w_v , z(u ) with certain combinatorial coefficients @xmath506 . computing @xmath507 , the derivatives can either all go on the integrand , giving [ terms1 ] _ w^dt_3^k f(u , t , t^2-w ) . or after @xmath508 derivatives on the integrand , the next derivative goes on the lower boundary , producing @xmath509 , and the remaining @xmath510 derivatives produce a sum of terms ( neglecting numerical coefficients for the moment ) [ terms2 ] w^-k+q+p2 + 12_2^p_3^q f(u , w,0 ) p+qk-1 . now , at @xmath85 , we have @xmath511 , hence the terms , inserted into become , respectively , [ terms ] ( uv)^2k-_u^dt_3^k f(u , t , t^2-u^2),(uv)^2k - u^-2k+2q+p+1_2^p_3^q f(u , u,0 ) . to obtain @xmath512 , these remain to be integrated with @xmath513 . the @xmath182-integrals are unproblematic at large @xmath182 by the falloff of the test function , but they may become singular at @xmath197 . the most singular terms are the latter ones in when @xmath514 , i.e. , @xmath515 . it is then obvious that the @xmath182-integrals over are finite multiples of @xmath516 , as long as @xmath517 . thus , for @xmath517 , @xmath518 is finite , and is in fact a polynomial in @xmath265 of degree @xmath519 , because @xmath520\leq k\leq \ell$ ] . if @xmath126 is an integer and @xmath521 , the most singular terms @xmath514 are _ 0^duu^n(uv - z)^2k - n-2w^-k+12 f(u , w,0)_w = w_v , z(u ) , with @xmath522\leq k\leq n+2 $ ] . while all other terms are finite multiples of @xmath523 at @xmath85 , these terms are logarithmically divergent at @xmath85 . to isolate the divergence , we split the integration range into the intervals @xmath524 and @xmath525 , for any fixed @xmath526 . the latter integral is a finite multiple of @xmath523 at @xmath85 . in the former , we write @xmath527 , so that the second contribution is also a finite multiple of @xmath523 at @xmath85 . the remaining terms , that diverge at @xmath197 when @xmath85 , are [ remaining ] f(0,0,0)_0^u duu^n(uv - z)^2k - n-2w_v , z(u)^-k+12 = : i^div_n , k(v , z ) . restoring the suppressed numerical coefficients , these terms sum up to [ sumup ] _ k=[2]^n+2 c^n+2_k 12 ( 12)_k-1 i^div_n , k(v , z ) = - f(0,0,0)_z^n+2 _ 0^u du u^n . comprises all contributions to @xmath528 that are at divergent at @xmath85 , while all other contributions are polynomials in @xmath265 of degree @xmath495 . the @xmath182-integral in can be performed explicitly : introducing the integration variable @xmath529 and the constant @xmath530 , the integrand is a linear combination of terms @xmath531 . if @xmath250 is odd , the primitive function is a polynomial in @xmath532 , @xmath533 , and @xmath534 . evaluated at the upper and lower values @xmath535 and @xmath536 , these are regular functions in @xmath265 and @xmath38 , that possess convergent power series expansions in @xmath537 and @xmath38 in the range @xmath538 . in particular , they contribute further finite values at @xmath85 to , that are polynomials in @xmath265 of degree @xmath495 . if @xmath539 is even , in addition to terms of the previous algebraic type , the primitive functions contain terms of the form a^2 + 2(s+)_s =- vz^u - vz = ( z^2-v^2z^2)^+1 . the logarithm of the numerator is again a convergent power series as above , and contributes further finite values at @xmath85 to , that are polynomials in @xmath265 of degree @xmath495 . but the denominator yields the logarithmic term @xmath540 . collecting all prefactors , we find the total logarithmic contribution to to be given by ( n+2 ) ! b_n(v)((1-v)z)f(0,0,0 ) with @xmath541 . with ( * ? ? ? * eqs . 15.3.19 , 15.3.5 ) , this can be brought into the manifestly polynomial form of @xmath542 as given in the lemma . knowing ( the form of ) the first @xmath495 derivatives of @xmath491 at @xmath85 , we obtain the claim of the lemma , for @xmath126 integer . if @xmath497+\varepsilon$ ] is not an integer , then all terms give rise to finite integrals @xmath543 as long as @xmath544 + 2 $ ] , i.e. , @xmath545 are polynomials in @xmath265 of degree @xmath519 up to @xmath546 + 2 $ ] . however , a scaling argument shows that @xmath547 + 2}_zi_n(v , z)(f)\vert_{z=0}$ ] has a subleading term of order @xmath548 : namely , the integrands g_k(u , v , z)=u^n(uv - z)^2k-[n]-2w_v , z(u)^-k+12 of the leading terms are homogeneous of order @xmath549 in @xmath182 and @xmath38 . using euler s equation in the form @xmath550 , this implies ( z_z-)_0^u dug_k(u , v , z ) = -ug_k(u , v , z ) , where @xmath551 - 2}u^\varepsilon$ ] . this differential equation for @xmath552 admits contributions @xmath553 with undetermined integration constants @xmath554 , that sum up to @xmath555 in the statement of the lemma . this proves the lemma for non - integer @xmath126 . the proof of the lemma clearly exhibits the origin of the logarithmic divergence to be the range @xmath556 of the integration over @xmath557 . notice also in the logarithmic singularity at @xmath558 , where @xmath559 . it arises upon integration over @xmath243 in the vicinity of @xmath38 , corresponding to the point @xmath367 . this singularity does not lead to divergences , because it is always tamed by the factor @xmath560 in @xmath561 . we work with the convention that the cut of @xmath562 ( @xmath563 ) is along @xmath564 $ ] . as usual we define _ 2(z):=-_c_zdz , _ 3(z):=_c_zdz , z[1 , ) , where @xmath565 is any smooth curve from @xmath352 to @xmath38 which does not intersect @xmath566 . with that @xmath567 and @xmath568 are analytic on @xmath569 . since ( -,0 ] u , [ 1,)u , the expression for @xmath414 is manifestly analytic for @xmath570 $ ] . the formula for @xmath414 can be derived by first computing the integral f(x)=^x dt(1-t^2)(q_1(t))^2 x , \|x\|>1 , which gives ( after analytic continuation to @xmath571 $ ] ) f(z)= ( ( ) ^2 - 3 + + ( 16 + 3 ) + 23 _ 2 + 2c_1 ) , where @xmath572 is an undetermined constant . a second integration yields @xmath414 for @xmath570 $ ] . here we use well - known identities for @xmath573 and @xmath574 ( see , e.g. , @xcite ) and 1u^2 - 1()^n-1= ( ) ^n ( n=2,3 ) , + 11u_2= _ 3 .the expressions on the l.h.s . are problematic , since they have poles at @xmath575 , which overlap with the cut along @xmath277 $ ] of the pertinent function in the numerator . but the boundary values at @xmath576 along both sides of the cut of the expressions on the r.h.s . are well defined distributions . to compute @xmath577 we use that the complex derivative is given by the infinitesimal differential quotient in any direction , in particular we may choose the direction of the real axis : f(z)_z = v+iw = f(v+iw ) and hence f(z)^z = v+i0_z = v - i0= ( f(v+i0)-f(v - i0 ) ) . in addition we give the following formulas : 1u^2 - 1^u = v+i0_u = v - i0&=&i((v+1)-(v-1 ) ) = i(1-\|v\| ) , v , + & = & -i ( 1-\|v\| ) , v , + _ 2(xi0)&=&-_0^x dt = ( x-1)ix , x , + _ 2(xi0)&=&_2 + 12 ( x)^2 + 6- x(x-1 ) , x>1 , + _ 3(xi0)&=&_0^x dt = ( x-1)2(x)^2 , x . with that the result is obtained by a straightforward calculation ( dropping terms involving @xmath578 ) . applying the functional @xmath579 = \int_{-1}^{+1 } dv \ ; \big[h_0(v ) - h_1(v)\partial_v\big ] f(v)$ ] ( with @xmath461 and @xmath421 as in prop . 4.2 ) to odd power functions @xmath580 , all integrals are of the types j_n & = & _ -1^+1dv v^n2 = ( -1)^n _ -1^+1dv v^n2 , + k_n & = & _ -1^+1dv v^n22 , + l_n & = & _ -1^+1dv v^n_22 = ( -1)^n _ -1^+1dv v^n_22 , so that [ hodd ] h_fish[v^2m+1 ] = 23 l_2m+1 + 43 j_2m+1 - ( 2m+1)(13 k_2 m -12 j_2m+1 ) -9 - 56 . since we could not find these integrals in the literature , we sketch their computation here . in @xmath438 , we partially integrate @xmath581 with primitive @xmath582 . this gives @xmath583 and the recursion j_n & = & - - nn+1 j_n-1 which is solved by [ in ] j_n= 2_=0^[n2]12 + 1 . summing the geometric series in the integrand of @xmath438 , we also get [ sumin ] _ n=0^j_n= _ -1^+1 2 = _ 2 2_v=-1^v=+1 = -6 . @xmath584 vanish if @xmath126 is odd . partially integrating @xmath585 in @xmath586 , expanding @xmath587 as a geometric series , and using , we get [ kn ] k_2 m = 22m+1_n=2m+1^j_n= ( 6 + _ n=0 ^ 2 m j_n ) . the integrals @xmath588 can be obtained by partial integration of the factor @xmath589 with primitive @xmath590 , which yields @xmath591 and the recursion ( n+1)l_n = 3 -(-1)^nn(j_n+j_n-1 ) - -n l_n-1 with solution [ ln ] ( n+1)l_n= 6 -(-1)^n_=n+1^j_= ( 1+(-1)^n)6 + ( -1)^n_=0^n j_. inserting , , into proves prop . 4.3 . * acknowledgments . * md profitted from discussions with gnter scharf and raymond stora during an early stage of this work . extensive discussions with klaus fredenhagen clarified many conceptual issues . we thank the anonymous referee for insisting , by his very detailed and qualified inquiries , on more detailed explanations in sect . [ sec : renorm ] , and for raising the interesting issue of the structure of the ope . 99 m. abramovitz , i.a . stegun : handbook of mathematical functions , dover publications , new york , 1972 . avis , c.j . isham , d. storey , _ quantum field theory in anti - de sitter space - time _ * d 18 * ( 1978 ) 35653576 . m. bertola , j. bros , u. moschella , r. schaeffer , _ a general construction of conformal field theories from scalar anti - de sitter quantum field theories _ , nucl . b * 587 * ( 2000 ) 619644 [ = arxiv : hep - th/9908140 ] ; + m. bertola , j. bros , v. gorini , u. moschella and r. schaeffer , _ decomposing quantum fields on branes _ , nucl . phys . b * 581 * ( 2000 ) 575603 . j. bros , h. epstein , u. moschella , _ towards a general theory of quantized fields on the anti - de sitter space - time _ , commun . * 231 * ( 2002 ) 481528 . r. brunetti , k. fredenhagen , _ microlocal analysis and interacting quantum field theories : renormalization on physical backgrounds _ , commun . * 208 * ( 2000 ) 623661 . m. dtsch and k. fredenhagen , _ a local ( perturbative ) construction of observables in gauge theories : the example of qed _ , commun . * 203 * ( 1999 ) 71105 . m. dtsch and k. fredenhagen , _ algebraic quantum field theory , perturbation theory , and the loop expansion _ , commun . * 219 * ( 2001 ) 530 . m. dtsch and k. fredenhagen , _ causal perturbation theory in terms of retarded products , and a proof of the action ward identity _ , rev . math . * 16 * ( 2004 ) 12911348 [ = arxiv : hep - th/0403213 ] . m. dtsch , k .- h . rehren , _ a comment on the dual field in the ads - cft correspondence _ , lett . * 62 * ( 2002 ) 171184 . m. dtsch , k .- h . rehren , _ generalized free fields and the ads - cft correspondence _ , ann . henri poincar * 4 * ( 2003 ) 613635 . h. epstein , _ on the borchers class of a free field _ , nuovo cim . * 27 * ( 1963 ) 886893 . h. epstein , v. glaser , _ the role of locality in perturbation theory _ h. poincar a * 19 * ( 1973 ) 211295 . freedman , k. johnson , j.i . latorre , _ differential regularization and renormalization : a new method of calculation in quantum field theory _ , * b371 * ( 1992 ) 353414 . c. fronsdal , _ elementary particles in a curved space . ii _ , phys . rev . * d10 * ( 1974 ) 589598 . r. haag , d. kastler , _ an algebraic approach to quantum field theory _ , j. math . phys . * 5 * ( 1964 ) 848861 . s. hollands , r.m . wald , _ local wick polynomials and time - ordered products of quantum fields in curved spacetime _ , commun . * 223 * ( 2001 ) 289326 ; + s. hollands , r.m . wald , _ existence of local covariant time - ordered products of quantum fields in curved spacetime _ , commun . * 231 * ( 2002 ) 309345 . s. hollands , r.m . wald , _ conservation of the stress tensor in perturbative interacting quantum field theory in curved spacetimes _ , rev . * 17 * ( 2005 ) 227312 . g. kllen , _ formal integration of the equations of quantum theory in the heisenberg representation _ , ark . fysik * 2 * ( 1950 ) 371410 . l. lewin : polylogarithms and associated functions , elsevier north holland , 1981 . maldacena , _ the large @xmath592 limit of superconformal field theories and supergravity _ , adv . * 2 * ( 1998 ) 231252 . v. moretti , _ comments on the stress - energy tensor operator in curved spacetime _ , commun . math . * 232 * ( 2003 ) 189221 . rehren , _ algebraic holography _ , ann . henri poincar * 1 * ( 2000 ) 607623 ; + k .- h . rehren , _ local quantum observables in the ads - cft correspondence _ , phys . b * 493 * ( 2000 ) 383388 . rehren , _ qft lectures on ads - cft _ , proceedings of the 3rd summer school in modern mathematical physics , zlatibor , serbia ( 2004 ) , b. dragovich ( ed . ) , belgrade 2005 , pp . 95118 [ = arxiv : hep - th/0411086 ] . w. rhl , _ lifting a conformal field theory from @xmath593-dimensional flat space to @xmath594-dimensional ads space _ , nucl . * b705 * ( 2005 ) 437456 . watson : a treatise on the theory of bessel functions , cambridge univ . press , 1958 ( @xmath595 edition ) . e. witten , _ anti - de sitter space and holography _ , adv . * 2 * ( 1998 ) 253291 .	the problem of perturbative breakdown of conformal symmetry can be avoided , if a conformally covariant quantum field @xmath0 on @xmath1-dimensional minkowski spacetime is viewed as the boundary limit of a quantum field @xmath2 on @xmath3-dimensional anti - desitter spacetime ( ads ) . we study the boundary limit in renormalized perturbation theory with polynomial interactions in ads , and point out the differences as compared to renormalization directly on the boundary . in particular , provided the limit exists , there is no conformal anomaly . we compute explicitly the one - loop `` fish diagram '' on @xmath4 by differential renormalization , and calculate the anomalous dimension of the composite boundary field @xmath5 with bulk interaction @xmath6 . # 1#1 -10 mm @xmath7 institut fr theoretische physik , universitt gttingen , + friedrich - hund - platz 1 , 37077 gttingen , germany + @xmath8 courant research centre `` higher order structures in mathematics '' , universitt gttingen , + bunsenstr . 35 , 37073 gttingen , germany _ dedicated to raymond stora on the occasion of his 80th birthday _ pacs 2001 : 11.10.-z , 04.62.+v msc 2000 : 81t05 , 81t10 , 81t20
equilibrium phase transitions are determined as non - analyticities of the free energy and can strictly appear only in the thermodynamic limit @xcite . at finite temperatures , phase transitions are driven by thermal noise and states of systems are minima of the free energy , namely systems tend to show large entropy for higher temperatures . on the other hand , quantum phase transitions are driven by quantum fluctuations and appear at absolute zero temperature @xcite . while the entropy contribution to the free energy in equilibrium systems prohibits the equilibrium phase transition in one - dimensional systems with short range interactions @xcite , a nonequilibrium phase transition in one dimension is possible . nonequilibrium phase transitions are usually considered as qualitative changes of the steady state . in classical case nonequilibrium phase transitions are well studied and appear e.g. in driven diffusive models @xcite . the yang - lee description of equilibrium phase transition in terms of zeros of the partition function @xcite , has been applied to classical nonequilibrium setting @xcite . on the other hand , nonequilibrium quantum phase transitions ( nqpt ) are much less known . in that respect some interesting numerical results @xcite show a nqpt in one - dimensional boundary driven nonequilibrium spin systems . moreover , approaches to deal with the nonequilibrium quantum systems usually involve different approximations , and their validity near a nqpt has not been carefully discussed . thus , it is important to analyze the differences between several approaches to nonequilibrium quantum systems in a simple pedagogic model undergoing a nqpt . nonequilibrium steady states ( ness ) of quantum systems are mainly studied using two approaches . * in one approach , we decompose the total , composite system into a finite system and reservoir parts and derive , by tracing out the latter , an effective master equation for the density operator of the former , i.e. we obtain an effective master equation for the reduced density operator of the finite system . it is possible to derive an exact master equation for the reduced density operator of the system , which however , is usually as difficult as to solve the original problem . therefore , various approximation schemes have been developed to suitably describe different regimes . in the simplest case , we obtain a markovian completely - positive master equation , namely the lindblad master equation @xcite , and the ness is obtained as a projection of the initial density operator onto the null - space of the generator of the dynamics , the liouvillian @xmath6 . markovian master equations have mostly been used in quantum optics @xcite , quantum information , and quantum computation as a simple model of noisy channels . however , recently they have also been applied in condensed matter context to study high temperature transport properties of simple one - dimensional systems @xcite and to discuss special nonequilibrium states of matter @xcite . in this paper , we shall employ two models , which have been developed to study transport of open quantum systems , namely the modified redfield master equation approach @xcite and the mesoreservoir approach @xcite . * the other is the @xmath0 algebra approach , which was first introduced with the purpose to rigorously formulate equilibrium statistical mechanics @xcite . it has been applied to infinitely extended systems @xcite . starting from ruelle s work @xcite on scattering - theoretical characterizations of ness , the @xmath0 algebra method has been extensively developed ( see for instance , references of @xcite ) and applied also in the context of phase transitions @xcite . contrary to the reduced density operator method , the @xmath0 algebra deals with infinite systems which evolve unitarily . it was shown that if a system is connected to two separated subsystems initially in thermal states , finite part of the total system approaches to a unique ness under some mathematical conditions @xcite . in that case the constructed ness does not depend on the initial decomposition , i.e. the size of the central ( finite ) system , and initial density operator of the central system . although the @xmath0 algebra provides mathematically rigorous results , the applications are quite limited . accordingly , it is beneficial to use the @xmath0 algebra approach to test other approaches for their validity . in this paper we use the @xmath0 algebra approach , the lindblad formalism with mesoreservoirs , and the modified redfield master equation to study the xy spin @xmath7 chain a paradigmatic model exhibiting quantum phase transitions ( qpt ) @xcite . the hamiltonian of the transverse - field xy spin 1/2 chain is @xmath8 where @xmath3 denotes the anisotropy , @xmath4 denotes the magnetic field , and @xmath9 are the pauli spin operators at @xmath10-th site . to begin with , let us summarize phase transitions of the xy spin @xmath7 chain discussed previously . at @xmath5 , there is a second order qpt characterized by the order parameter @xmath11 ( @xmath12 for @xmath13 and @xmath14 for @xmath15 ) , which separates the ferromagnetic ordered phase ( @xmath16 ) and the paramagnetic disordered phase ( @xmath17 ) . equilibrium average of @xmath18 for @xmath19 is always zero , and it can not be used as an order parameter . magnetization in the @xmath20 direction is always finite , but susceptibility is divergent at absolute zero temperature and @xmath5 . in addition , at @xmath21 and @xmath22 there is a quantum phase transition between the ordered phase in the @xmath23 direction ( @xmath13 ) and the @xmath24 direction ( @xmath15 ) . in case of the open xy spin 1/2 chain coupled to local lindblad reservoirs @xcite and redfield reservoirs @xcite , indications of a nqpt were reported . numerically , it was observed that @xmath25 is a critical magnetic field which separates phases showing different scaling of the quantum mutual information ( qmi ) , spectral gap , and far - from - diagonal spin - spin correlations @xmath26 ( @xmath27 ) . this transition was recently described by an information - geometric approach using a fidelity distance measure @xcite . we shall complement the so far obtained phase diagram of the xy chain by using the @xmath0 algebra method . we shall focus on the nqpt and reveal new quantum phase transitions , which do not exist in equilibrium . we shall show the existence of a nqpt at @xmath5 for arbitrary temperatures of reservoirs , which is associated to a discontinuity of the third derivative of the off - diagonal elements of the correlation matrix . we also demonstrate a discontinuity of the first derivative of the correlation matrix elements at @xmath28 . in addition , we show that qpt at @xmath5 persists in the nonequilibrium case if temperature of at least one reservoir is set to be zero . a phase diagram showing equilibrium and nonequilibrium quantum phase transitions of the xy spin 1/2 chain is presented in . we shall compare nqpts obtained using the @xmath0 algebra method with those obtained by simulations of the reduced density operator approaches , namely the mesoreservoir method @xcite and the modified redfield master equation @xcite . we shall also discuss the thermalization properties of the mesoreservoir approach . the rest of the paper is structured as follows . in section 2 we apply the @xmath0 algebra approach and show analytically the four different nonequilibrium quantum phase transitions of the xy model . in section 3 we discuss the master equation approaches . first , in subsection 3.1 , we present the results of the lindblad mesoreservoir approach and discuss its equilibrium and nonequilibrium properties . second , in subsection 3.2 , we study the modified redfield model . we discuss the results and conclude in section 4 . in this section we study the nqpts of the xy spin 1/2 chain using the results of the @xmath0 algebra approach . we shall show the coexistence of two nqpts at @xmath5 and @xmath29 ( the critical xy line ) and a nqpt at the critical magnetic field @xmath30 . we also discuss the phase transition at the critical point @xmath21 and @xmath5 . these results are contrasted with two reduced density operator approaches , which will be discussed in . the xy spin 1/2 model has been extensively studied with the aid of the @xmath0 algebra , however the nonequilibrium quantum phase transitions have not been discussed so far . we briefly explain the terminology of the @xmath0 algebra method in the [ app2 ] . in @xcite the ness of the model was rigorously constructed using the scattering theory proposed by ruelle @xcite . long range correlation @xcite and non - negativity of entropy production @xcite have been discussed for this ness . those works are highly mathematically oriented . in contrast , we provide a simpler and direct calculation , and focus on the nonequilibrium phase transitions . we start with an infinite chain , which is separated into the left semi - infinite part ( @xmath31 ) , the central finite part ( @xmath32 ) , and the right semi - infinite part ( @xmath33 ) . the infinite chain is initially in the product states @xmath34 where @xmath35 is an arbitrary state and @xmath36 is the density operator of the canonical ensembles with temperature @xmath37 ( @xmath38 ) . one can prove that a unique ness of the chain exists and that the diagonal modes with positive ( negative ) velocities are distributed according to the canonical ensemble of the left ( right ) reservoir , and satisfy the wick s theorem , which is explicitly expressed as eqs . and . let us first diagonalize the xy spin 1/2 model . with the aid of the araki - jordan - wigner transformation ( see the [ app2 ] ) , the hamiltonian is mapped to a chain with fermions @xmath39 where @xmath40 satisfies canonical anti - commutation relations : @xmath41 next we apply the fourier transformation @xmath42 and rewrite the hamiltonian in the momentum basis as @xmath43 the hamiltonian can easily be diagonalized @xmath44 where eigenmode annihilation operators @xmath45 and the energy @xmath46 are defined by @xmath47 the ness of this model uniquely exists and is fully characterized by @xcite @xmath48 and the wick s theorem @xcite , where @xmath49 represents a ness average , @xmath50 are fermi distributions with temperatures @xmath51 , @xmath52 is a step function , and @xmath53 is a velocity @xmath54 . to be more precise , modes with positive and negative velocity follow different kms conditions ( see eq . ( 27 ) of @xcite for the xy chain . see also eq . ( 54 ) in @xcite for xx chain , where a more explicit expression in terms of the femi distribution is provided in eq . ( 65 ) . ) . let us briefly discuss the intuitive idea behind equations and , which is implemented rigorously with the @xmath0 algebra approach . for that sake note that in the diagonal form the hamiltonian can be interpreted as a sum of hamiltonians @xmath55 and @xmath56 of two noninteracting systems where @xmath57 and @xmath58 . consider a free system such as one described by @xmath59 . at @xmath60 it is split into three parts : a left semi - infinite chain , a central finite chain , and a right semi - infinite chain . the left ( right ) semi - infinite chain is in a thermal state with temperature @xmath61 ( @xmath62 ) . at @xmath63 the three pieces are connected and the particles ( or quantum waves described by the diagonal modes ) that where confined to the left can now propagate to the right without any scattering and the same holds for the particles at the right . in that way , every right going mode in the system ( @xmath64 ) comes from the left and is populated accordingly with @xmath65 and every left going mode in the systems ( @xmath66 ) comes from the right and is populated by @xmath67 . this is the content of the equation . the equation is the same but for the system described by @xmath68 , i.e. , @xmath69 and the energies and velocities replaced according to @xmath70 using , and changing variables back to the @xmath40 , we compute the two point correlation functions in the ness @xmath71 from above equations , it is clear that the real parts of nonequilibrium correlations @xmath72 and @xmath73 are simply the averages of equilibrium correlations at the temperatures of left and right reservoirs . on the other hand , the imaginary parts of @xmath72 are zero , whereas the imaginary parts of @xmath74 are non - zero only in nonequilibrium and hence possess purely nonequilibrium features of the steady state , e.g. the information about the heat current . let us first discuss the magnetization , that is known to have a second order quantum phase transition ( for instance see proof of theorem 2.3 . in @xcite ) @xmath75 where @xmath76 is the susceptibility in ness . we shall now show that this transition persists also in the nonequilibrium setting where one of the reservoirs is initiated at finite non - zero temperature . from equation , one can see that the magnetization @xmath77 does not depend on the spatial variable @xmath10 , and both the magnetization and susceptibility are simply the average of those in equilibrium . due to @xmath78 , the susceptibility diverges if and only if at least one of the reservoirs has absolute zero temperature . the divergence with respect to temperature @xmath51 and the difference of magnetic field from one @xmath79 is logarithmic . therefore , at @xmath5 we have a quantum phase transition and a nqpt associated with the divergence of the susceptibility . the mechanism of the nqpt associated to the logarithmic divergence of the magnetic susceptibility is the same as in the equilibrium case since the ness average of the magnetization is a sum of terms coming from left and right reservoirs in equilibrium . since all real parts of the correlation matrix have this property , a genuine nqpt should be discussed through @xmath80 , which vanishes for equilibrium state ( the kernel is proportional to the difference of fermi distributions calculated with the initial temperatures of the left and right semi - infinite parts . ) . the first derivative of @xmath81 reads @xmath82 the presence of the step function @xmath83 in produces a discontinuity at @xmath28 . physically it can be explained by a change of the direction ( sign of velocity ) of the diagonal modes @xmath45 as a function of the magnetic field @xmath4 . in the equilibrium case , diagonal modes @xmath45 follow the same fermi distribution independent of the direction ( velocity ) of the modes . on the other hand , in ness , diagonal modes @xmath45 with positive and negative velocities follow different fermi distributions . therefore , changing the direction of the modes alter their nature only in nonequilibrium and this is the origin of the nqpt at @xmath84 , which is consequently a genuine nonequilibrium phenomenon . interestingly , there is also a discontinuity in the third derivative at @xmath5 . the third derivative for @xmath85 reads @xmath86 from eq . ( [ eq : third_derivative ] ) , one can evaluate the jump of the third derivative at @xmath5 : @xmath87 since this transition is not related to the change of the direction of the modes , its origin is different from the transition discussed for @xmath28 . the behavior of derivatives and is depicted in . it is interesting that the non - continuities of the derivatives are present also in case of nonzero temperatures of the initial states of the reservoirs @xmath88 , while the transition of a real part such as magnetization is possible iff at least one of the reservoirs is set to absolute zero temperature . thus , behavior of the imaginary part is quite different from the real part , which is also the reason we believe @xmath89 determine genuine nonequilibrium properties . we recall that imaginary parts are related to currents of conserved quantities , e.g. the energy current and the spin current ( the latter is well - defined only in the isotropic case ) . at @xmath21 and @xmath90 we find a square root divergence of magnetic susceptibility with the temperature @xmath37 and the difference of magnetic field from one @xmath79 , namely @xmath91 . in this case @xmath92 , the finite temperature nonequilibrium transitions associated to discontinuities in the derivatives of the imaginary parts of the correlation functions disappear . equilibrium and nonequilibrium phase diagram of the xy spin 1/2 model is depicted in . finally , we discuss scaling of the quantum mutual information ( qmi ) in the ness of the @xmath0 algebra . the scaling of qmi has been recently discussed in the context of area laws . in @xcite it was shown that gaussian thermal states obey the area law for qmi . on the other hand , the steady states may violate the area law despite being gaussian . in fact qmi was used to characterize the nqpt at @xmath28 in the xy model with redfield reservoirs @xcite ( see ) . in @xcite it was shown that below the critical field ( @xmath93 ) qmi obeys the area law , whereas above the critical field the qmi scales linearly with the system size . we calculate the scaling of the qmi in the ness by numerically diagonalizing the two - point majorana correlation matrix of the @xmath0 algebra . the von neumann entropy of the block of size @xmath94 can be expressed in terms of the eigenvalues @xmath95 of the @xmath96 two - point majorana correlation matrix ( defined in ) as @xcite @xmath97 the quantum mutual information of the block of size @xmath98 is then given by @xmath99 . in we show that the qmi scales logarithmically with the block size @xmath94 . interestingly this scaling does not depend on the system parameters and therefore does not capture the nonequilibrium phase transitions . a related analytical calculation of the area law violation in the @xmath0 algebra ness was recently reported in the free fermion case @xcite , which is equivalent to the isotropic limit of the xy spin 1/2 chain . in this section we compare obtained exact @xmath0 algebra results with two models of open quantum systems , namely the lindblad mesoreservoir and the modified redfield master equation . in both cases dynamics of the system is determined by a liouville equation @xmath100 + { { \hat{\mathcal{d}}}}\rho,\end{aligned}\ ] ] where @xmath101 $ ] denotes the commutator . although exact forms of the hamiltonian @xmath102 and the dissipator @xmath103 depend on details of the model , in the studied cases the hamiltonian is quadratic @xmath104 and the lindblad ( coupling ) operators are linear in terms of majorana fermions : @xmath105 therefore , as has been shown in @xcite , ness is a gaussian state and is fully characterized by the @xmath106 majorana correlation matrix @xmath107 of which time evolution is determined by @xmath108 where @xmath109 and @xmath110 denote real and imaginary parts of the reservoir matrix @xmath111 , which represents influence of the environment and depends on the details of the dissipator for the mesoreservoir case is given in [ app1 ] and for the modified redfield case along with the derivation in @xcite . ] . time derivative of the correlation matrix is zero in ness , hence the steady state correlation matrix @xmath112 is obtained as a solution of a lyapunov equation @xmath113 the existence of ness guarantees that the equation ( [ eq : lyap ] ) has a solution , which can be computed efficiently in @xmath114 steps . recently , a model to describe ness of the systems including finite reservoirs degrees of freedom was introduced @xcite . the idea is not to trace out all degrees of freedom of the reservoirs but rather keep some representative parts ( mesoreservoirs ) , which are physically interpreted as the contacts of the system with the reservoirs . the time evolution of the total system ( mesoreservoirs and the system of interest ) is modeled with the lindblad master equation such that the mesoreservoirs are thermalized if the couplings between the system and mesoreservoirs are zero . in @xcite it was shown that in the limit of weak coupling and large mesoreservoir size the system is thermalized if two mesoreservoirs with same thermodynamic variables are attached , while the particle current follows the landauer formula if two reservoirs with different thermodynamic variables are attached . therefore , one can argue that the mesoreservoir approach gives a meaningful description of a thermal reservoir . in the previous work , particle conserved hamiltonian attached to a linear dispersion mesoreservoirs are discussed . although the linear dispersion mesoreservoir gives the same liouvillian spectrum ( see [ app1 ] ) , their thermalization properties can be different . to mimic the @xmath0 algebra setup discussed in the previous section , we treat parts of a chain as mesoreservoirs . the total hamiltonian consists of one dimensional xy spin 1/2 chain with a system size @xmath115 : @xmath116 where @xmath117 is given in ( [ eq : xy ] ) . we interpret parts of the system @xmath118 $ ] and @xmath119 $ ] as mesoreservoirs , and the remaining part @xmath120 $ ] as a central system . mesoreservoir parts ( @xmath121 ) are thermalized using the lindblad dissipator @xmath122 where @xmath123 and @xmath124 are diagonal modes of @xmath125 and @xmath126 , respectively . the operators @xmath127 and @xmath128 can be interpreted as couplings between mesoreservoirs and the traced - out degrees of freedom ( super - reservoir ) . in the subsection [ sec : meso1 ] we first study the equilibrium properties of the model . in particular , we will discuss the divergence of magnetic susceptibility @xmath129 at @xmath5 . then , in the subsection [ sec : meso2 ] , we turn to the nonequilibrium quantum phase transitions obtained by the @xmath0 algebra . in this subsection we first study magnetization and the corresponding susceptibility in the equilibrium state by the lindblad mesoreservoir approach . we observe a highly fluctuating magnetization profile for @xmath130 , and a flat magnetization profile except at the boundaries between system and mesoreservoirs for @xmath85 . typical magnetization profiles for different @xmath131 are shown in ( red dashed lines indicate the @xmath0 algebra results ) . the fluctuations present for magnetic field @xmath132 are suppressed by decreasing dissipator strength @xmath131 . moreover , the spin profiles approach the @xmath0 results ( red ) by decreasing @xmath131 ( see ) . because of the fluctuations with respect to a space variable @xmath10 , we define magnetization and susceptibility using a space average of @xmath133 . one can see in _ left _ that the magnetization is roughly reproduced . however , there are small fluctuations for @xmath130 . this fluctuations are clearly observed in the susceptibility shown in _ right_. in particular , one can see a big jump in susceptibility at the critical point @xmath84 . on the other hand , for @xmath85 , the lindblad approach agrees very well with the @xmath0 algebra . in ( a ) we show the numerically calculated correlations @xmath134 in the mesoreservoir approach and compare them with the exact analytic @xmath0 results . near - diagonal correlations at low temperatures ( @xmath135 ) calculated with the mesoreservoir approach agree with the @xmath0 results . on the other hand , long - range correlations saturate as a function of @xmath136 ( for finite @xmath131 and @xmath137 ) to a plateau in the mesoreservoir approach , while they decay exponentially fast in the @xmath0 algebra method . as expected , the agreement between the @xmath0 and mesoreservoir correlations is improved for smaller coupling to the super - reservoir @xmath131 and larger mesoreservoir size @xmath137 ( see ( b ) ) . if the temperature is high enough @xmath138 the mesoreservoirs fail to thermalize the central system , as can be seen from ( c),(d ) , where we show the @xmath131 dependence of @xmath139 for low and high temperatures at different values of the magnetic field @xmath4 . the mesoreservoir results match with the @xmath0 algebra for low temperatures , but are notably different for high temperatures . fortunately in the regime where we want to observe the qpt , the @xmath0 algebra and the lindblad mesoreservoir approach agree well . nevertheless , the thermalization at @xmath5 is very subtle since the susceptibility diverges logarithmically . to discuss the divergence at @xmath5 one should take a limit @xmath140 and @xmath141 , as can be seen from figure [ kdependence ] , where we show the mesoreservoir size @xmath137 dependence of the susceptibilities with several coupling strengths @xmath131 at @xmath142 . we see that the divergence of susceptibility for @xmath143 is reconstructed with the lindblad mesoreservoir approach in the limit of @xmath140 and @xmath141 . the divergence is logarithmic with respect to @xmath137 and @xmath37 . imaginary parts of correlation matrix are vanishing linearly with respect to @xmath131 , while specific observables such as heat current vanish for arbitrary @xmath131 . in this subsection , we shall discuss the ness of mesoreservoir approach . we numerically observe that the real part of the correlation matrix is an average of equilibrium correlation matrices . following the discussion of the previous subsection , the divergence of susceptibility is concluded in the mesoreservoir approach . because of this property , we focus on the imaginary part of the correlation matrix , which represents a genuine nonequilibrium property . let us first discuss the two extreme cases ( @xmath144 and @xmath145 ) . for very small @xmath131 the total system ( system @xmath146 reservoirs ) is decoupled from super reservoirs , and the system is expected to follow an average of fermi distributions @xmath147 . thus , @xmath148 and @xmath149 approach to zero for @xmath144 . on the other hand , taking large @xmath131 enforces the mesoreservoirs to follow exactly the fermi distributions , as a result , @xmath148 and @xmath149 are very small for @xmath145 ( having non - zero off - diagonal element is related to the deviations of reservoirs particle occupations from the fermi distributions @xcite ) . therefore , one should carefully choose coupling @xmath131 to deal with the imaginary parts of nonequilibrium case , and in fact , @xmath131 should be in the order of @xmath150 to describe coherent transport in ness , where @xmath151 is the width of the mesoreservoirs energy band . in ( a ) and ( b ) show comparisons between the ness correlations of the @xmath0 algebra and the mesoreservoir at different magnetic fields ( a ) and different dissipation strengths @xmath131 . we observe that the correlations of the @xmath0 algebra and the mesoreservoir in ness quickly differ with the distance from the diagonal elements . despite the excellent agreement for nearly diagonal elements , we were unable to obtain the jumps in derivatives of the correlation function . moreover , we have to choose the right ratio between the coupling strength @xmath131 and the mesoreservoir size @xmath137 as can be seen from ( c ) and ( d ) , where we show the off - diagonal correlations for different coupling strengths and a fixed mesoreservoir size @xmath152 . since the ratio between @xmath131 and @xmath137 is important for the nonequilibrium correlations , one sees the strong @xmath137 dependence on the correlations , contrasting the fact that real parts of the correlations do not drastically depend on the mesoreservoir size @xmath137 . for instance , real parts shown in are quite robust by changing @xmath137 ( thus , we show only @xmath153 ) , on the other hand , clearly shows that imaginary parts strongly depend on @xmath137 . moreover , as discussed in @xcite , the @xmath131 dependence of @xmath148 shows a plateau for @xmath154 , where @xmath155 is a monotonic decreasing function of @xmath137 . in conclusion , similar to the boundary driven lindblad model we observe also in the mesoreservoir approach signatures of the nqpt at @xmath84 , namely a change in the sensitivity of local observables to the system parameters . below the critical field @xmath28 we see large fluctuations , whereas above the critical field there are no fluctuations as we vary the model parameters ( see and ) . however , this fluctuations are suppressed as @xmath131 goes to zero . on the other hand , we were unable to observe the scaling of the qmi and the far - from - diagonal correlations due to limitation of the resources , since we should take the limit of large mesoreservoir size and small coupling to the super - reservoir . nevertheless , the behavior of the off - diagonal correlations shown in suggests that the mesoreservoir approach in the limit @xmath156 and @xmath157 resembles the behavior of the @xmath0 algebra , i.e. power law decay of the ness correlations and the logarithmic divergence of the qmi in all regimes . we also recover the @xmath0 algebra divergence of the magnetic susceptibility at zero temperature and @xmath5 in equilibrium and nonequilibrium situations . other genuine nonequilibrium transitions observed with the @xmath0 algebra could not be recovered in the mesoreservoir approach . the modified redfield master equation was studied in @xcite as a model of thermal reservoirs . the main difference to the lindblad reservoirs is the non - local property of the redfield dissipator which , when extending the integrals in the correlation function from minus infinity to infinity , ensures that the gibbs state is the steady state if all reservoirs have the same temperature . after using this additional assumption , the redfield dissipator takes the form @xmath158 + h.c . , \label{eq : redfd}\ ] ] where @xmath159 is the reservoir correlation function , @xmath160 are the coupling operators , and the tilde @xmath161 denotes the interaction picture . the reservoir correlation function @xmath162 satisfies the kms condition @xmath163 we follow @xcite and couple the xy spin 1/2 chain to two boundary thermal reservoirs with the coupling operators @xmath164 the left and right reservoirs are uncorrelated @xmath165 and have ohmic reservoir spectral functions @xmath166 where @xmath167 is a fourier transform of the correlation function @xmath168 . this choice of correlation functions and coupling operators represents standard reservoirs of harmonic oscillators at two ends with possibly different temperatures . notice , that due to the extension of lower integral bound to minus infinity in the dissipator ( [ eq : redfd ] ) the frequency cutoff in the spectral functions is irrelevant . in this paper we use the following parameters : @xmath169 for @xmath170 . in case of equal temperatures and large system sizes , therefore we have @xmath171 is equal to @xmath94 . ] @xmath172 , we necessarily recover the results of the @xmath0 algebra approach for all values of the anisotropy @xmath3 and magnetic field @xmath4 ( see _ left _ ) . however , in the nonequilibrium setting , where one temperature remains constant and the other goes to zero , the magnetic susceptibility remains finite even for large spin chains and @xmath5 ( see _ right _ ) . we also compare the @xmath0 algebra and redfield ness expectation values of the magnetization for different values of the magnetic field , and observe disagreement around @xmath5 ( see _ left _ ) . in _ right _ we show numerically computed susceptibility . we observe large fluctuations below the critical magnetic field @xmath28 . this large fluctuations are in agreement with previously observed fluctuations of local observables and the characterization of the long - range magnetic correlation phase with hypersensitivity of local observables on model parameters @xcite . for a detailed discussion of the nqpt at @xmath28 in the xy spin 1/2 chain using the master equation approach , see @xcite . here we also compare the nonequilibrium susceptibility at @xmath5 , which in contrast to the @xmath0 result remains finite even if one temperature of the reservoirs goes to zero ( see _ right _ and _ right _ ) . further we note that the discontinuities of the first and the third derivative of @xmath89 at @xmath28 and @xmath5 were not reproduced using the redfield mesoreservoir approach . moreover , we find thatthe imaginary part of the correlation matrix strongly depend on the dissipation strength @xmath131 . ( from bright to dark ) . _ left _ : equilibrium susceptibility ( @xmath173 ) , the dashed line corresponds to the @xmath0 algebra . _ right _ : nonequilibrium susceptibility with @xmath174 and @xmath175 . the red dashed line denotes the @xmath0 algebra result in nonequilibrium . other parameters : @xmath176.,title="fig : " ] ( from bright to dark ) . _ left _ : equilibrium susceptibility ( @xmath173 ) , the dashed line corresponds to the @xmath0 algebra . _ right _ : nonequilibrium susceptibility with @xmath174 and @xmath175 . the red dashed line denotes the @xmath0 algebra result in nonequilibrium . other parameters : @xmath176.,title="fig : " ] algebra . the redfield and the @xmath0 algebra results disagree if the difference of temperatures is large . the dashed line corresponds to the @xmath0 algebra , and the grey lines from bright to dark correspond to redfield calculations for different systems sizes @xmath177 . other parameters : @xmath178 , @xmath179 . , title="fig : " ] algebra . the redfield and the @xmath0 algebra results disagree if the difference of temperatures is large . the dashed line corresponds to the @xmath0 algebra , and the grey lines from bright to dark correspond to redfield calculations for different systems sizes @xmath177 . other parameters : @xmath178 , @xmath179 . , title="fig : " ] we studied nonequilibrium quantum phase transitions in the xy spin 1/2 chain analytically using the @xmath0 algebra method . first , we showed that the qpt at @xmath5 is present also in nonequilibrium if one temperature of the reservoirs remains at absolute zero . in other words , qpt persists even with strong thermal noise coming from one of the reservoirs if the other reservoir is at absolute zero temperature . at the critical point @xmath21 and @xmath5 the logarithmic divergence of susceptibility becomes algebraic . second , we discovered two new transitions which do not exist in equilibrium state . to be concrete , we found discontinuities of the first derivative ( at @xmath28 , @xmath180 ) and the third derivative ( at @xmath5 , @xmath180 ) of the imaginary part of the correlation matrix . the former transition appear because a part of normal modes changes its sign of velocity when the magnetic field is smaller than @xmath181 , and thus those modes change the information of the reservoirs they carry . on the other hand , the physical interpretation of the later transition is still unclear . moreover , at the critical point @xmath21 and @xmath5 the jumps in the derivatives disappear . we use these transitions to test the utility of two time generators commonly used in theories of the reduced density operator , namely lindblad and redfield master equations . * _ comparison with the lindblad mesoreservoir approach _ we show that the lindblad mesoreservoir quantitatively reproduces the qpt in equilibrium . however , we observed that off - diagonal elements which have small expectation values disagree even in equilibrium case . for nonequilibrium state , we numerically observe that the real part of correlation is an average of equilibrium values , which agrees with the @xmath0 algebra , and does not agree with the modified redfield equation . + for the correlation matrix , we found that the lindblad mesoreservoir and the @xmath0 algebra agree only for correlations near diagonal elements , and are more accurate for small magnetic fields and low temperatures . despite the good agreement for nearly - diagonal elements , we were not able to recover nonequilibrium phase transitions except the divergence of the susceptibility at @xmath5 , which seems to be induced by the same mechanism as in equilibrium case . this may be the effect of finite mesoreservoir size @xmath137 and coupling strength to environment @xmath131 . * _ comparison with the modified redfield master equation _ the modified redfield equation by construction exactly describes equilibrium states , which is described by the @xmath0 algebra in the thermodynamic limit . nevertheless , it does not reproduce any nonequilibrium phase transitions observed by the @xmath0 algebra . moreover , we find that the imaginary part of the correlation matrix strongly depend on the dissipation strength @xmath131 . for the reduced density operator methods ( redfield and lindblad ) , the hypersensitivities to the model parameters below @xmath181 are reported . in the mesoreservoir case the fluctuations are suppressed if small dissipator strength @xmath131 and large mesoreservoir size @xmath137 are taken . thus , a drastic change of system s properties at @xmath84 is a common feature of the @xmath0 algebra and the reduced density operator methods , but they are quite different . in the previous works , the transition obtained by reduced density operator was characterized by the appearance of the correlation resonances @xcite and different scaling of the qmi in the long and short range correlation regimes . on the contrary , it was shown with the @xmath0 algebra that the scaling of the correlation elements with the distance from the diagonal remains unchanged as we cross the critical magnetic field @xmath181 , i.e. , exponential decay of correlations in the equilibrium and power law decay of correlations in the nonequilibrium case . we also numerically showed that the qmi scales logarithmically with the system size @xmath94 in all regimes . therefore , we conclude that the transition obtained by the reduced density operator @xcite is a consequence of the approach itself . since none of the discussed reduced density operator approaches can describe all transitions obtained by the exact calculations ( @xmath0 algebra ) , the question `` can any reduced density operator methods thermalize the xy spin 1/2 chain in the complete range of parameters , and at the same time reproduce the nonequilibrium phase transitions obtained by the @xmath0 algebra ? '' remains open . further another interesting question arises , namely `` what kind of nonequilibrium quantum phase transitions we obtain by using different approaches ? '' . in other words , to what extend the nonequilibrium quantum phase transitions and the nonequilibrium properties of systems in general depend on the reservoirs ( models of open system evolution ) . authors are grateful to toma prosen and shuichi tasaki for valuable discussion . b acknowledges the fondecyt grant 3130495 . sa acknowledges the fondecyt grant 3120254 . fb acknowledges the fondecyt grant 1110144 and anr - conicyt 38 grant . in this appendix , we give the form of the time evolution of the correlation matrix for the lindblad case . then , we derive the time evolution equation in the case of xy 1/2 chain with mesoreservoirs . the liouvillian of our interest has the following form , @xmath182+\sum_\mu ( 2l_\mu\rho l_\mu^\dag - l_\mu^\dag l_\mu\rho -\rho l_\mu^\dag l_\mu ) \\ h & = & ( w_1 , \cdots , w_{2n } ) { \bf h } ( w_1,\cdots , w_{2n})^t \\ l_\mu&=&\sum_j l_{\mu , j}w_j\ , \end{aligned}\ ] ] where @xmath183 is the majorana operator satisfying @xmath184 , and @xmath185 is an anti - symmetric matrix . following @xcite , we attach a hilbert space structure on a linear @xmath186 space of operators acting on a conventional hilbert space , where the inner product of the hilbert space is defined by the hilbert - schmidt norm , i.e. , @xmath187 next , linear maps @xmath188 and @xmath189 over the hilbert space are defined by @xmath190 prosen showed that the number @xmath191 is conserved for the time generation of the given liouvillian @xcite , and the unitary time evolution and dissipator for a space spanned by the basis @xmath192 with even @xmath193 are given by @xmath194 \rangle } } & = & -i4h_{j , k } \hat{c}^\dag_j \hat{c}_k { { \vert \rho \rangle}}\ , \\ { { \vert \hat { { \cal d } } \rho \rangle } } & = & \sum_\mu { { \vert 2l_\mu\rho l_\mu^\dag - l_\mu^\dag l_\mu\rho -\rho l_\mu^\dag l_\mu \rangle } } \\ & = & \sum_{j , k } \left ( 4 m_{j , k } c_j^\dag c_k^\dag -2 ( m_{j , k}+m_{j , k } ) c_j^\dag c_k \right){{\vert \rho \rangle}}\ , \\ m_{j , k } & \equiv & \sum_\mu l_{\mu , j } l_{\mu , k}^ * = \sum_\mu \vec{l}_\mu \vec{l}^\dag_\mu\ .\end{aligned}\ ] ] hereafter , we shall focus on the dynamics of the space with even @xmath193 so that the liouvillian has the following form @xmath195 where @xmath196 and @xmath110 are defined by @xmath197 using the above form , the time evolution of the correlation matrix @xmath198 can be discussed . thanks to @xmath199 , the correlation matrix reads @xmath200 thus , the time evolution of the anti - symmetric part of the correlation matrix @xmath201 follows @xmath202 it gives the matrix form presented in : @xmath203 therefore ness averages of correlation functions are given by the lyapunov equation : @xmath204 let us write down matrices @xmath185 and @xmath111 in terms of majorana operators . first , matrix @xmath185 is given by @xmath205 where matrices @xmath206 and @xmath207 are defined by @xmath208 next , we give the matrix @xmath111 . let @xmath209 be a diagonal mode of the mesoreservoir parts ( @xmath210l , r ) @xmath211 then , the matrix @xmath111 reads @xmath212 where matrix @xmath213 is defined as follows @xmath214 where @xmath215 and @xmath216 following @xcite , we summarize the setup of the @xmath0 algebra approach and the araki - jordan - wigner transformation . the norm - completion of the algebra generated by the pauli spin matrices forms the @xmath0 algebra @xmath228 , and the infinite extension of the hamiltonian defines a @xmath0 dynamical system whose dynamics is given by a group of strong continuous * -isomorphism which is formally given by @xmath229 . in this approach states of the system @xmath230 are represented by positive functionals over the @xmath0 algebra , and the hilbert space is introduced as a representation of ( @xmath231 ) . physically speaking , a @xmath0 algebra @xmath217 represents a set of observables with finite expectation values , a group of strong continuous * -isomorphism @xmath232 gives a time evolution over the @xmath0 algebra , and a state @xmath233 gives a correspondence between observables and expectation values @xmath234 . the equilibrium states with a given @xmath235 and at temperatures @xmath236 are defined as the states @xmath237 satisfying the kms condition : @xmath238 as we explained in the main part , a system is initially decomposed into three parts ( left semi - infinite , finite system , and right semi - infinite parts ) . then , the initial condition is given by @xmath239 where @xmath240 are the kms states at temperatures @xmath241 ( @xmath242=l , r ) of left and right distinct part respectively . the existence of the unique ness associated to this initial condition @xmath243 was proved in @xcite . moreover , tasaki et . al . proved that the ness is independent of the initial partitions , and is stable against local perturbation @xcite . the ness average we study in the main part is a ness with respect to @xmath244 . next , let us review the araki jordan wigner transformation of the spin chain . let @xmath245 be the @xmath0 algebra generated by @xmath246 , @xmath247 , and @xmath248 which satisfy @xmath249 let @xmath250 be the @xmath0 algebra generated by @xmath245 and an element @xmath236 satisfying @xmath251 where @xmath252 is an automorphism of @xmath245 which satisfies @xmath253 then , @xmath228 is the subalgebra of @xmath250 : @xmath254 where @xmath255 it gives a fermionic hamiltonian in the main part . in the main part , we denote @xmath256 for the involution @xmath257 to give a physical presentation . the hermitian conjugate of operators acting on hilbert space satisfies the definition of the involution .	we study nonequilibrium quantum phase transitions in xy spin 1/2 chain using the @xmath0 algebra . we show that the well - known quantum phase transition at magnetic field @xmath1 persists also in the nonequilibrium setting as long as one of the reservoirs is set to absolute zero temperature . in addition , we find nonequilibrium phase transitions associated to imaginary part of the correlation matrix for any two different temperatures of the reservoirs at @xmath1 and @xmath2 , where @xmath3 is the anisotropy and @xmath4 the magnetic field strength . in particular , two nonequilibrium quantum phase transitions coexist at @xmath5 . in addition we also study the quantum mutual information in all regimes and find a logarithmic correction of the area law in the nonequilibrium steady state independent of the system parameters . we use these nonequilibrium phase transitions to test the utility of two models of reduced density operator , namely lindblad mesoreservoir and modified redfield equation . we show that the nonequilibrium quantum phase transition at @xmath1 related to the divergence of magnetic susceptibility is recovered in the mesoreservoir approach , whereas it is not recovered using the redfield master equation formalism . however none of the reduced density operator approaches could recover all the transitions observed by the @xmath0 algebra . we also study thermalization properties of the mesoreservoir approach .
inflation is the standard paradigm for the generation of perturbations in matter density that produces large - scale structure ( lss ) in the universe @xcite . although the inflationary paradigm has successfully predicted various properties of the observable universe , including flatness and a nearly scale - invariant spectrum of perturbations @xcite , the correct model of inflation has yet to be confirmed . the simplest inflation models predict nearly gaussian primordial perturbations , though more complex models such as multi - field inflation @xcite posit a departure from a gaussian distribution . alternatives to inflation , such as the ekpyrotic model of cyclic expansion and contraction @xcite , also predict non - gaussian primordial perturbations . since a detection of non - gaussianity would discriminate between these fundamentally different models , much work is being done to constrain non - gaussianity , both in lss through the galaxy distribution and through anisotropies in the cosmic microwave background ( cmb ) . primordial non - gaussianity is readily probed through measurements of the bispectrum of the cmb , in which a nonzero measurement constitutes a `` smoking gun '' detection , modulo any systematic effects . some alternative probes of non - gaussianity in lss include the galaxy bispectrum , which is plagued by nonlinearities , and galaxy cluster abundances and dark - matter halo clustering , which suffer from low - number statistics . the accepted parametrization of primordial non - gaussianity is to introduce a quadratic term to the primordial potential @xmath5 , written as @xmath6 where @xmath7 is a gaussian random field @xcite . this form describes local - type non - gaussianity with an amplitude @xmath2 . the latest constraint on @xmath2 is from the wilkinson microwave anisotropy probe s ( wmap ) seven - year bispectrum , which gives @xmath8 at 95% c.l . @xcite , which will soon give results from its first data release , is expected to produce constraints on @xmath2 of order @xmath9 @xcite . one useful effect of non - gaussianity that has gained much attention , and which we explore in this analysis , is a distinct scale - dependent clustering bias on large scales @xcite . specifically , it has been shown that @xmath2-type non - gaussianity produces a shift in the bias that behaves as @xmath10 ; this can suppress or enhance clustering on the largest scales . various authors have used the large - scale angular power spectra of lss tracers to constrain @xmath2 assuming a scale - dependent bias . @xcite , using the sloan digital sky survey ( sdss ) @xcite data release 5 ( dr5 ) @xcite , derived the constraints @xmath11 at 95% c.l . using the photometric quasar sample and @xmath12 at 95% c.l when these quasars are combined with other data sets , such as the integrated sachs - wolfe effect ( isw ) and luminous red galaxies ( lrgs ) . this analysis was extended in @xcite to include the sdss dr6 @xcite photometric quasar catalog compiled by @xcite , hereafter rqcat . @xcite extended this analysis to a two - parameter curvaton model @xcite while others have sought to use this method in combination with other data sets @xcite . @xcite also showed that _ planck _ and _ euclid _ @xcite together could possibly detect @xmath13 . finally , work on using multiple tracers of different bias @xcite has predicted future limits of @xmath14 . as in all auto - spectrum measurements , particularly on large angular scales , potential correlations caused by systematic effects must be removed to isolate the correlation signal from the lss tracer . systematics have been a concern in recent analyses of the _ baryon oscillation spectroscopic survey _ ( boss ) @xcite . studies of boss photometric galaxies @xcite attempted to remove systematic effects by fitting an amplitude for each systematic template , and subtracting its estimated contribution from the power spectrum . recent work in @xcite outlines a formalism for how photometric calibration variations affect power spectrum and cosmological parameter estimations . another method for removing systematics is mode projection , which explicitly marginalizes over the amplitude of each template that traces the spatial dependence of a potential systematic error . this method has been profitably used for foreground removal in the low-@xmath15 cmb power spectrum ( e.g. @xcite ) . this method is more robust than subtracting the estimated contamination from the power spectrum estimator , because it is _ exactly _ insensitive to any systematic that traces the spatial dependence of the template . it also avoids the possibility of `` oversubtraction '' the phenomenon in which chance correlations of the real galaxy density with systematics templates lead to an overestimate of the contamination , and hence lead to a downward bias in the low-@xmath15 power spectrum . in this paper , we attempt to remove large - scale systematics from the latest photometric quasar sample available , which is rqcat . we investigate the photometric quasar sample because , by probing large redshifts , quasars are able to probe scale - dependent bias more effectively than other matter tracers , as demonstrated in @xcite . we use photometric quasars from a larger redshift range than this previous analysis , using redshifts as low as @xmath16 and as high as @xmath17 . we divide this sample into three redshift slices ( 0.91.3 , 1.62.0 , 2.32.9 ) and construct three angular cross - power spectra to test if they are consistent with zero , which we expect since quasars at different redshifts should be uncorrelated on large scales . we also construct templates of various potential systematic effects and try mode - projecting them from the angular cross - power spectrum estimator . note that the sample used by e.g. @xcite was constructed by a different algorithm and is _ not _ the restriction of the rqcat sample to an earlier , smaller footprint . therefore the specific amplitude of systematic errors herein should not be taken as quantitatively indicative of the level of contamination in these previous analyses . we find that without mode - projecting any systematics , the first and second redshift slices are correlated with an 9.2@xmath1 significance , and the first and third redshift slices are correlated with a 2.7@xmath1 significance , both on large scales . this implies that systematics are heavily contaminating large - scale correlations in the quasar maps . using mode projection , we find significant reductions in the correlations between redshift slices after removing templates corresponding to the stellar density , stellar locus offsets , airmass , and seeing . we are able to decrease the significance of the correlation between the first and third redshift slices to 1.8@xmath1 ; however , the significance of the correlation between the first and second redshift slices decreased only to 5.4@xmath1 , which is indicative of significant contamination . while the quasar catalog exhibits @xmath32% contamination , measuring @xmath4 will require a contamination level less than 1% . we conclude that other systematic effects are still present that are contaminating the sdss photometric quasar sample . we must improve methods of construct cleaner photometric quasar samples if we hope to use them to provide competitive constraints on @xmath4 on large scales . the plan of our paper is as follows : in sec . [ s : data ] we describe the photometric quasar sample we use in the analysis . in sec . [ s : theory ] we describe briefly the theory behind scale - dependent bias and how it affects the angular power spectrum . in sec . [ s : method ] , we outline the estimators we use , the method of mode projection , and the systematics for which we search . in sec . [ s : results ] , we give the angular cross - power spectrum measurements , before and after mode projection , and we discuss its implications . we conclude in sec . [ s : conclude ] . wherever not explicitly mentioned , we assume for the background cosmology and power spectrum a flat @xmath18cdm cosmology with parameters compatible with the wmap7 data release @xcite . we use the photometric quasars from the sdss dr6 @xcite to trace the matter density and construct its angular spectrum . the sdss consists of a 2.5-m telescope @xcite with a 5-filter ( _ ugriz _ ) imaging camera @xcite and a spectrograph . automated pipelines are responsible for the astrometric solution @xcite and photometric calibration @xcite . bright galaxies , luminous red galaxies ( lrgs ) , and quasars are selected for follow - up spectroscopy @xcite . the data used here were acquired between august 1998 and june 2006 and are included in sdss data release 6 @xcite . specifically , we use the photometric quasar catalog ( rqcat ) composed by @xcite . the entire catalog consists of 1,172,157 objects from 8417 deg@xmath19 on the sky selected as quasars from the sdss dr6 photometric imaging data . quasars are the brightest objects at large redshifts ( @xmath20 ) , making them better tracers of the matter density at large scales than lrgs . we limit our dataset in this analysis to uv - excess ( @xmath21 ) quasars . specifically , we implement this choice by requiring the catalog columns * good * @xmath22 and * uvxts * = 1 . we divide quasars into 3 redshift slices @xmath23 , and @xmath24 . we plot the three redshift distributions in fig . [ f : redplot1 ] , while their properties are given in table [ t : redshifts1 ] . the procedure for constructing the redshift distributions is described in the appendix and is similar to that described in ref . @xcite . for the survey geometry we use the dr6 survey mask as a union of the survey runs retrieved from the sdss cas server . we omitted runs 2189 and 2190 because many objects in these runs were cut from rqcat . this mask was pixelized using the mangle software @xcite . we pixelize the quasars as a number overdensity , @xmath25 , onto a healpix pixelization @xcite of the sphere with @xmath26 , where @xmath27 is the pixel s number of quasars divided by the pixel s survey coverage @xmath28 and @xmath29 . we try to clean our sample by removing pixels that we suspect are contaminated . we reject pixels with extinction @xmath30 , full widths at half - maximum of its point - spread function ( psf ) fwhm @xmath31 arcsec , and stellar densities ( smoothed with a @xmath32 fwhm gaussian profile ) @xmath33 stars / deg@xmath19 ( twice the average stellar density ) , similar to @xcite . the extinction cut is very important because a high extinction affects the @xmath34 band , which is crucial to identifying quasars . also , since stars tend to be misidentified as quasars , it seems prudent to cut regions with high stellar density . we implement these cuts using dust maps from @xcite and a stellar overdensity map , smoothed with a 2@xmath35 fwhm gaussian profile , constructed from stars ( @xmath36 ) with @xmath37 ( cutoff for point sources in rqcat ) from the sdss dr6 @xcite . we also reject pixels for which the survey region covers less than 80% of the pixel area . we also excised several rectangular regions that appeared to contain missing data ; the angular rectangles in equatorial ( j2000 ) coordinates that were removed are @xmath38 , @xmath39 , @xmath40 , @xmath41 , @xmath42 , @xmath43 , @xmath44 , @xmath45 , + @xmath46 , and @xmath47 . finally , we cut regions comprising healpix @xmath48 pixels that contained no stars for which to estimate the stellar color locus and flux - error locus offset systematic templates , which use photometric measurements of stars , as well as cut pixels with color locus offsets greater than 5 mag ( see sec . [ s : system ] ) . after these cuts , the survey region comprises 122,780 pixels covering a solid angle of @xmath49 deg@xmath19 . the quasar maps for each slice are shown in fig . [ f : qsomaps ] . note that we reduce a large - scale systematic by implementing an additional cut on the quasars in the highest redshift bin to exhibit a high kernel - density - estimator ( kde ) quasar probability density , requiring the catalog column * qsodens * @xmath22 ( see fig . [ f : qsomaps ] ) . the derivation for the scale - dependent bias model for non - gaussianity of the `` local '' ( @xmath4 ) type is detailed in @xcite , for which we give a brief overview . the model assumes the halo model scenario ( see @xcite for a detailed review ) in which all matter is contained in a distribution of halos on large scales which mimics the distribution of matter on small scales . in the gaussian case , fluctuations on small and large scales are uncorrelated . however , a signature of local - type non - gaussianity is that the @xmath50 term in the gravitational potential ( see eq . [ e : fnl1 ] ) causes small - scale matter fluctuations to correlate with large - scale halo fluctuations due to mixing of their respective gravity perturbations . positive ( negative ) @xmath2 incurs a positive ( negative ) correlation between scales and an increase ( decrease ) in the halo bias on large scales as compared to the gaussian case . specifically , this scale - dependent shift in the bias was derived for a general halo mass function @xmath51 in @xcite to be @xmath52 where @xmath53 is the wavenumber corresponding to the length scale probed , @xmath54 is the transfer function , @xmath55 is the the growth factor normalized such that @xmath56 , @xmath57 is the speed of light , @xmath58 is the hubble parameter today , and @xmath59 is the rms overdensity in a sphere of radius @xmath60 mpc . for the case of a universal mass function , such as the press - schechter @xcite or sheth - tormen @xcite mass functions , this expression was shown in @xcite to reduce to @xmath61 the expression first derived in @xcite , where @xmath62 is the matter density today relative to the critical density for a flat universe , and @xmath63 is the critical density of spherical collapse . the parameter @xmath64 ranges from 1 for lrgs , which populate all halos equally , to 1.6 for quasars that populate only recently merged halos . we use eq . ( [ e : dbuni ] ) with @xmath65 in our analysis to model the scale - dependent bias of quasars . the clustering bias is a key component of the matter power spectrum @xmath66 , thus we expect scale - dependent bias to affect it as well as the angular power spectrum @xmath67 . since we are interested in large correlations in this analysis , we can not use the small - scale limber approximation , but must use the full expression for @xmath67 including redshift - space distortions . these expressions are given for the scale - independent bias case in eqs . 29 and 30 in @xcite and for the scale - dependent bias case in @xcite . we assume the linear matter power spectrum in our analysis mpc@xmath68 . for our redshifts , this corresponds to @xmath69 . we only use @xmath67 for @xmath70 , so it is safe to use the linear matter power spectrum in our analysis . ] . the @xmath67s were calculated for all 3 redshift slices and presented in fig . [ f : cl ] , including for the cases of nonzero @xmath4 . non - gaussianity , which produces a shift in the bias of the form @xmath10 , manifests itself on large angular scales . if any systematic effect can increase power on large scales , we will be unable to distinguish it from scale - dependent bias . one way to attempt to assess the contamination of our quasar maps would be to cross - correlate the quasar maps with each of the systematics and see if it is nonzero , which was done in @xcite with stars and red stars . however , this will be difficult to do for the other systematics we consider because the errors are uncertain . thus , in this analysis we compare angular cross - power spectra before and after mode - projecting an individual systematic to see if it decreases . more specifically , we first calculate the angular cross - power spectrum without attempting to remove any systematics . we then recalculate it , except that we mode - project one systematic . finally , we compare the two to see if the spectra are significantly different . since there are three redshift bins , there will be 3 cross - spectra . the cross - spectra are useful because we expect different quasar maps to be uncorrelated if they are clean since objects from separate redshift bins should not correlate with each other . and [ f : clqcrsys ] , where we see a finite cross - correlation on smaller scales . ] however , if the same systematic effect appears in two quasar maps , then their cross - spectra will be non - zero . so our goal for each cross - spectra is to identify those systematics for which mode - projection decreases the magnitude of the cross - spectra , and then try mode - projecting this set of systematics together to see if the cross - spectra goes to zero , which would be the limit of the usefulness of this test . if it remains non - zero , then we know that there remains systematic effects in the maps that are unidentified . of course , this will only identify those systematics which contaminate two or all three quasar maps . there may be systematic effects that contaminate only one of the maps that would not appear in the cross - spectra yet still contaminate any estimates of @xmath4 using auto - spectra . we decide not to search for this case since the auto - spectra we would need for the test would also be used to to constrain @xmath4 . however , this case should be considered in future work . in order to construct our angular power spectrum estimator , we begin by constructing the data vector @xmath71 where @xmath72 and @xmath73 length-@xmath74 vectors constituting the two maps we are cross - correlating , making @xmath75 have a length of @xmath76 . the covariance matrix for @xmath75 is given by @xmath77 where @xmath78 and @xmath79 ( @xmath80 and @xmath81 ) are the signal ( noise ) covariance matrices for map 1 and map 2 , respectively , and we neglect any cross - correlation between the maps . @xmath82 is given by @xmath83\ , , \end{aligned}\ ] ] where @xmath67 is the theoretical angular power spectrum ( we assume @xmath84 for our prior ) , @xmath85 is a legendre polynomial and @xmath86 is the pixel window function , and we assume poisson noise for @xmath87 given by @xmath88 in terms of the mean number of galaxies per pixel @xmath89 . note that we project out the monopole and dipole of the power spectra since these moments are not of interest cosmologically . the simplest way to remove a data point is to set its error to infinity ( or a large value , so 1/error@xmath99 ) . since the errors for multipoles with low shot noise are proportional to the multipole , setting the monopole and dipole to large values actually removes their estimator contributions to @xmath100 . giving large values to these moments increases their covariances in the estimator , making their contributions negligible , while increasing estimate errors slightly . this method is a specific case of the more general projection method we use throughout the paper ( see sec . [ s : modeproj ] ) . this method was first presented in @xcite , and was also presented in @xcite and was applied to cmb foreground removal by @xcite . note this is not equivalent to constraining the monopole and dipole moments to large values . this is performed in our analysis by setting the @xmath67 prior equal to 100 for @xmath101 and 1 . our treatment of masked pixels is similar . we construct our estimator in real space , such that masking pixels , or striking missing pixels from the covariance matrix before inversion , is equivalent to setting the noise in these pixels to infinity , which removes their effect from the estimator anyway . it is true that masking causes the @xmath67 measurements to be correlated , but these correlations are described by the fisher matrix @xmath102 , such that @xmath103 in the estimator expressed in eq . [ e : fq ] removes the leakage , as shown in sec . 3e of @xcite . we use the method of mode projection to remove modes due to systematics from our angular power spectrum estimators , instead of subtracting them directly , which could lead to spurious modes remaining . this method projects out systematics by in effect giving these modes large noise values so that the estimator is rendered insensitive to them . mode projection @xcite has been used often for cmb data ( see @xcite for a review ) , yet this method can also clean lss analyses . we begin with the assumption that the systematic s effect on the qso map @xmath75 is of the form @xmath104 where the vector @xmath105 is the template for systematic @xmath106 . we then project out the systematics by adding to the covariance matrix @xmath107 in both the quad vector and fisher matrix ( see eq . [ e : fq ] ) the outer product of each systematic template with itself , according to @xmath108 which yields an unbiased estimator when @xmath109 . we set @xmath110 for each template large enough such that the resulting @xmath111 and @xmath112 converge . in this section we enumerate the various systematics templates for which we search in rqcat . the systematics we consider are extinction , stellar contamination , red stellar contamination , stellar color locus offset , stellar flux - error locus offset , air mass , seeing , sky brightness , modified julian date , camera column , and atmospheric refraction . we describe the possible effects of each systematic below , as well as the construction of the templates . we list the template numerical properties , including the template mean , standard deviation , @xmath113 , and @xmath114 vs. @xmath115 slopes , in table [ t : sys ] . for air mass , seeing , sky brightness , mjd , camera column , and atmospheric refraction , the required values were given per sdss scanning field , where each field is a spherical rectangle on the sky . we used the mangle @xcite software to convert these fields to healpix pixels . also , note that each template had its mean subtracted . _ extinction ( ext ) : _ the galactic extinction was taken from the infrared dust maps described in @xcite and given in terms of the reddening @xmath116 in magnitudes . the quasar catalog can be sensitive to any errors in the normalization or shape of the extinction law @xmath117 . also , the extinction correction is sensitive to the intra - band spectral energy distributions ( sed ) of the source under consideration ; quasars often have extreme intraband colors and this could lead to a variation of the quasar density determined from `` extinction - corrected '' photometry with @xmath116 . _ stellar contamination ( star ) : _ we use the stellar overdensity map mentioned in sec . [ s : data ] as the template . when constructing a quasar catalog , the main challenge is to separate stars from quasars , since both appear as point sources . the stellar contamination rate for rqcat is @xmath33% at lower redshifts , yet more precision may be required for non - gaussianity measurements . _ red stellar contamination ( rstar ) : _ the template is similar to stellar contamination , except the stellar sample is narrowed further to include only those that also satisfy the color cut @xmath118 . note that this template will also project our contamination from blue stars , since the `` blue star '' map is a linear combination of `` stars '' and `` red stars . '' _ stellar color locus offset ( ugr , gri , riz ) : _ this is sensitive to position - dependent offsets in the photometric calibration of the survey . we can measure these variations through shifts in the stellar locus , or scatter plots of the stars between two colors . templates were constructed , in terms of the ordinate vs. abscissa of the plot , for color loci @xmath119 vs. @xmath120 ( @xmath121 ) , @xmath122 vs. @xmath119 ( @xmath123 ) , and @xmath124 vs. @xmath122 ( @xmath125 ) in magnitudes . we start by determining the color locus for the sample of stars used in our stellar overdensity map , which we then fit to a line . we then divide the sample into 49,152 healpix @xmath48 pixels , with each large pixel containing 16 @xmath126 small pixels we use in our analysis so that there are sufficient stars to measure the color locus offset . in each large pixel , we use its stars to construct a color locus for that pixel . we then take the median of the difference between each star s ordinate in the large pixel s color locus with the linear fit of the entire sample s color locus as a measure of the offset . we then give the offset for all 16 small pixels the value of the offset calculated from the large pixel . _ stellar flux - error locus offset ( uerr ) : _ statistical photometric ( flux ) errors can scatter objects into or out of the region of color space used for quasar selection ( or the region of color space that maps to a particular photo-@xmath127 range ) , and thus affect the density of selected quasars . variations of these errors across the sky can then produce spurious large scale power in the quasar maps . a template was used only for the @xmath34 band . we construct the template in a manner similar to the stellar color locus offset template , except that ( 1 ) the ordinate and abscissa for the stellar loci are the flux error and flux ( in magnitudes ) , respectively , and ( 2 ) instead of a linear fit , we just fit the stellar locus to a horizontal line ( which is the same as averaging the ordinate values ) . _ air mass ( airu ) : _ air mass is the path length of atmosphere a photon travels from its source to the telescope , normalized to unity at zenith ; it depends on both sky position and observation time . increased airmass increases both the statistical errors on the magnitudes of objects , as well as producing color terms since the bluest wavelengths are selectively attenuated . the 5 filters in sdss are taken in rapid succession , so taking the air mass value from the @xmath34 band observation should be sufficient for template projection purposes . the values for the template are taken from the catalog archive server ( cas ) @xcite . _ seeing ( seeu , seer ) : _ seeing is defined here as the fwhm of the point spread function ( psf ) in a pixel , given in arcseconds from the sdss cas . it affects the photometric noise , de - blending of sources , and star - galaxy ( i.e. point source - extended source ) separation . templates were constructed for the @xmath34 ( @xmath128 ) and @xmath129 ( @xmath130 ) bands . _ sky brightness ( skyu , skyi ) : _ sky brightness is namely the brightness of the atmosphere during observation . small - scale variations in the sky brightness can affect object identifications , in that the sky brightness is subtracted from the bands before identification is performed . the values , given in maggies / arcsec@xmath19 , are taken from the sdss cas . templates were constructed for the @xmath34 band ( @xmath131 ) , in which sky brightness is dominated by moonlight if present , and the @xmath106 band ( @xmath132 ) , in which sky brightness is dominated by light from the intrinsic atmosphere . _ modified julian date ( mjd ) : _ the modified julian date ( mjd ) is a linear unit of time , given in days , to track astronomical observations . the mjd values for the sdss scanning fields were taken from the sdss cas . mode - projecting this template will project out modes from any systematic that varies linearly with observation time . _ camera column ( cam1 ... cam5 ) : _ the ccd camera for sdss consists of six ( 6 ) camera columns ( camcols ) . the ccds are highly calibrated ; however , an unknown malfunction in a camcol could misidentify objects , contaminating the quasar density field . the camcols for the sdss scanning fields are given in the sdss cas . a template for each camcol is _ almost _ binary , in that for camcol @xmath133 , the template value for a pixel is 1 if the pixel was observed by camcol @xmath133 and 0 if it was not . the `` almost '' is because the healpix pixels cross the scanning field boundaries , causing cases where a pixel is observed by multiple fields . in these cases the template value for camcol @xmath133 is given a fractional value denoting the percentage _ of the observed part _ of the pixel that was observed by camcol @xmath133 . for example , lets say a pixel was observed by two fields with different camcols , where camcols 2 and 4 observed 60% and 35% of the pixel , respectively , and 5% of the pixel was either not observed at all or observed by a bad field which had to be culled from the data . in this case , the templates for camcols 2 and 4 would have values 60/(60 + 35)=0.63 and 35/(60 + 35)=0.37 , respectively . it is easy to see that the sum of all the camcol templates must be a vector with all ones , which we call @xmath134 in this paper is _ not _ equal to the identity matrix . ] . thus , the template for camcol 6 is linearly dependent on the templates for camcols 15 , making it unable to remove contaminated modes that are independent of the other camcols . therefore we only mode - project camcols 15 . we also orthogonalize these templates to speed the convergence of the estimator . _ atmospheric refraction ( ref ) : _ refraction of photons through the atmosphere can perturb the position of a source , distorting the density field and increasing / decreasing correlations . since refraction is chromatic , the different bands will give different positions to a source . the template is designed to measure the elongation of the refraction - induced `` rainbow '' in the scan direction . the template is given in terms of the altitude ( @xmath135 ) and position angle ( @xmath7 ) of the telescope as @xmath136 , which is proportional to the quadrupole field that describes the perturbations due to refraction . the @xmath135 and @xmath7 values for each scanning field were taken from the sdss data archive server ( das ) . while each of the systematic templates has a `` primary motivation , '' each one may be sensitive to several contaminating effects . for example , the stellar color locus test is obviously sensitive to photometric calibration errors , but also contains some sensitivity to galactic dust and to the variation of stellar populations over the sky . the stellar density maps were motivated by the possibility of stellar contamination , however they are also correlated with the dust map since both are largest at low galactic latitudes . however , each of these effects is independent of the true quasar density and correlations of the quasar map with any of them would be problematic ; note also that since the template projection removes any linear combination of the input templates , there is no requirement that these templates be orthogonal , or sensitive to only a single root cause of systematic error . we also list the pearson correlation coefficients for all the templates in table [ t : corr ] . we see that almost all the templates have low correlations with other templates . of the 171 possible template pairings , only 7 of them have correlations ( positive or negative ) greater than @xmath13720% . the only two templates correlated with each other by more than 50% are the seeing templates in the @xmath34 and @xmath129 bands ( 80% ) . this implies that mode - projecting this set of the templates should remove a diverse set of modes . note that if the range of variation in conditions is large ( e.g. as for the airmass or sky brightness ) , the spurious contribution to the quasar density may be nonlinear in the template e.g. we may have @xmath138 where @xmath139 is a template ( e.g. the airmass ) and @xmath140 is a nonlinearity coefficient . since the template projection method is a linear method , we can only remove the @xmath140-term by adding yet another template proportional to @xmath141 ; in general , if multiple templates have large variations we would also have to include products of template maps . we tested the mode - projection of products of templates for the cross - correlation of z01 and z02 since it has the larger cross - correlation . the cross - correlation power spectra between the redshift slices before mode projection are shown in fig . [ f : qcross ] . out of 12 cross - power measurements ( 4 @xmath15-bins times 3 cross - correlations ) , two exhibit significant cross - powers . the largest one is between slices z01 and z02 , or @xmath142 , with a value of @xmath143 ( @xmath144 ) in the first @xmath15-bin ( @xmath145 ) . we see there is also a sizable cross - correlation between slices z01 and z03 , or @xmath146 , appearing in the second @xmath15-bin ( @xmath147 ) with a value of @xmath148 ( 2.7@xmath1 ) . large - scale biases are not expected to be this high , particularly with the magnitude of @xmath142 . this leads us to suspect that systematic errors in these slices are corrupting these quasar maps . the redshift slices z02 and z03 exhibit no significant cross - correlation . after mode - projecting each of our candidate systematic templates , we find several templates that significantly increased the cross - correlations . we list these templates for both cross - correlations in tables [ t : sysf1 ] and [ t : sysf2 ] . the biggest contributors by far to @xmath142 seem to be seeing in the @xmath34 and @xmath129 bands . mode - projecting each of these individually reduce the cross - correlation significance by more than 3@xmath1 . red star contamination , photometric calibration errors , and air mass also seem to play a role . our uncovering of red star contamination seems to be in agreement with @xcite , which found both of their samples of quasars to be contaminated by them . mode - projecting all these templates yields a cross - correlation measurement of @xmath149 ( @xmath150 ) , which is closest to the value from @xmath34-band seeing . it appears that seeing is a very important systematic effect in our quasar maps , and it should be monitored in future lss experiments . we tried to decrease @xmath142 further by mode - projecting products of templates . we tested several combinations , including ( @xmath151)@xmath19 , ( @xmath121)@xmath19 , ( @xmath152)@xmath19 , ( @xmath128)@xmath19 , ( @xmath130)@xmath19 , @xmath153 , @xmath154 , @xmath155 , @xmath156 , @xmath157 , @xmath158 , @xmath159 , @xmath160 , and @xmath161 none of the combinations we tested decreased the cross - correlation significantly . it is important to note that although we were able to reduce the cross - correlation significance by 3.8@xmath1 , the cross - correlation is far too high . we conclude that more systematics are present that were not found . for @xmath146 , stellar contamination is the dominant systematic . mode - projecting this template reduces the cross - correlation significance by more than half a @xmath1 , enough to drive the significance below 2@xmath1 . methods to reduce stellar contamination should also be developed for future lss surveys . seeing systematics seem to be having a small affect on this cross - correlation measurement as well . when we mode - project stellar contamination and seeing , we find a cross - correlation value of @xmath162 ( @xmath163 ) . final plots for @xmath142 and @xmath146 are shown in fig . [ f : clqcrsys ] . note we can use the measured cross - correlations to estimate the level of contamination in the redshift slices . for example , we can estimate the z01-z02 cross - correlation due to contaminants as @xmath164 , where @xmath165 is the fraction of contaminants . this implies that the contamination levels before mode projection for @xmath142 , @xmath146 , and @xmath166 in the relevant contaminated @xmath15-bins i simply use the first @xmath15 bin . ] are ( [email protected])% , ( [email protected])% , and @xmath1672.5% , respectively . after mode projection , the contamination levels for @xmath142 and @xmath146 reduce to ( [email protected])% and ( [email protected])% , respectively . from the theoretical auto - correlation plots in fig . [ f : cl ] , it appears we need the contamination level to be @xmath31% or less to see fluctuations from @xmath168 or 0.6% for @xmath169 . it is important to note that we can divide the systematic errors in our photometric quasar maps into two categories _ contamination _ and _ calibration_. stellar contamination is the principal example of the former since we are selecting point sources : it results due to the imperfection of the photometric redshift method in identifying quasars . ( in principle , crowding could lead to `` negative '' contamination since the presence of a higher stellar density may reduce the completeness of the catalog . ) @xcite claim a 96% efficiency in identifying uvx quasars with @xmath170 , yet this probably decreases for higher redshift quasars . this may explain why stellar contamination arises in the cross - correlation of z01z03 and not z01z02 . the other type of systematic error is calibration , where variations in the observing conditions or instrument properties propagate to changes in the completeness of the quasar catalog . this may consist of first - order calibration effects ( as in the case of photometric calibration errors , or airmass effects that shift the center of the @xmath34 band ) , as well as second - order or noise - induced effects such as that contributed by the sky brightness . although contamination has been the main focus of previous photometric catalogs , it appears that even the excellent internal calibration of the sdss does not provide samples with the uniformity on large scales needed for @xmath4 studies . we searched for various systematic effects in the sdss dr6 photometric quasar catalog @xcite , which is the latest catalog available . we find that the clustering signal from this catalog is contaminated on the largest scales . in particular , our z01 subsample ( @xmath171 ) and our z02 subsample ( @xmath172 ) exhibits a 9.2@xmath1 detection of cross correlation on scales @xmath173 , even though quasars from separate redshifts on these scales should be uncorrelated . we suspect that the cause of this large cross - correlation is systematic effects . by mode - projecting several candidate templates individually , we find that the maps are indeed contaminated by stars , photometric calibration errors , and variations in air mass and seeing , with seeing being the dominant contaminant . by mode - projecting this group of templates together , we were able to decrease the significance of the cross - correlation by more than 3@xmath1 . however , it is still too high to provide a competitive constraint on @xmath2 through scale - dependent bias . we find spurious number density fluctuations of @xmath174% rms , whereas we need a contamination level less than 1% ( 0.6% ) in order to measure values of @xmath4 less than 100 ( 10 ) . we recommend new efforts into improving photometry in order to use photometric quasars to constrain nongaussianity on large scales . we thank g. richards , s. ho and o. dor for useful discussions and comments . ap was supported during part of this research by doe de - fg03 - 92-er40701 , nasa nng05gf69 g , the gordon and betty moore foundation , and a nasa einstein probe mission study grant , `` the experimental probe of inflationary cosmology . '' ap is currently supported by an appointment to the nasa postdoctoral program at the jet propulsion laboratory , california institute of technology , administered by oak ridge associated universities through a contract with nasa . ch is supported by the us department of energy under contract de - fg03 - 02-er40701 , the alfred p sloan foundation , and the david and lucile packard foundation . funding for the sdss and sdss - ii has been provided by the alfred p. sloan foundation , the participating institutions , the national science foundation , the u.s . department of energy , the national aeronautics and space administration , the japanese monbukagakusho , the max planck society , and the higher education funding council for england . the sdss web site is http://www.sdss.org/. the sdss is managed by the astrophysical research consortium for the participating institutions . the participating institutions are the american museum of natural history , astrophysical institute potsdam , university of basel , university of cambridge , case western reserve university , university of chicago , drexel university , fermilab , the institute for advanced study , the japan participation group , johns hopkins university , the joint institute for nuclear astrophysics , the kavli institute for particle astrophysics and cosmology , the korean scientist group , the chinese academy of sciences ( lamost ) , los alamos national laboratory , the max - planck - institute for astronomy ( mpia ) , the max - planck - institute for astrophysics ( mpa ) , new mexico state university , ohio state university , university of pittsburgh , university of portsmouth , princeton university , the united states naval observatory , and the university of washington . all rights reserved . in @xcite , photometric redshift distributions @xmath175 for quasars were constructed using spectroscopic data from 2slaq @xcite . these redshift distributions determine how the matter overdensity @xmath176 relates to the quasar overdensity @xmath177 @xmath178dz\ , .\end{aligned}\ ] ] we refer the reader to @xcite for the theory behind this method , and we describe our method which is similar to and follows from @xcite . the expression for @xmath179 is given as @xmath180\pi_i(z')dz'\ , , \end{aligned}\ ] ] where @xmath181 is the linear bias as a function of redshift , @xmath182 is the comoving radial distance , and @xmath183 is the probability distribution for the quasar redshifts . the second term in eq . [ e : reddistbasic ] is due to magnification bias , which becomes important for large redshifts , with the lensing window function @xmath184 given for a flat universe by @xmath185\ , , \end{aligned}\ ] ] and @xmath186 being the logarithmic slope of the number counts of quasars as a function of flux : @xmath187 . in this appendix , we describe the method of calculating the various factors in eq . [ e : reddistbasic ] that comprise @xmath179 . we rely on spectroscopic data to compose redshift distributions . specifically , we use spectroscopic quasars ( spectro - qsos ) from an area with high spectroscopic coverage to construct a preliminary probability distribution @xmath188 and @xmath186 . we use spectroscopic data from the 2slaq survey , which contains 8389 spectro - qsos over its total region of view . we restrict ourselves to using spectroscopic data from five rectangles on the sky with declination range @xmath189 and right ascension ranges @xmath190 , @xmath191 , @xmath192 , @xmath193 , and @xmath194 , the same as those used in @xcite , for the quasar analysis . these rectangles in particular have high spectroscopic coverage and contain 5383 quasars . since we also need photo-@xmath127s for the quasars to construct probability densities , we only use quasars that have matches in the rqcat , decreasing the number of objects to 3443 . we calculate @xmath195 for each photometric redshift slice using a kernel density estimator of the form @xmath196 where @xmath197 is the number of spectro - qsos ( matched with a photo - qso in rqcat ) in the photometric redshift slice , @xmath198 is the spectro-@xmath127 of the @xmath53th - matched quasar , and @xmath1 is the slice s kernel width . @xmath1 is chosen to be smaller than any real features in @xmath188 yet large enough to smooth out shot noise . table [ t : prelimstat ] lists @xmath197 and @xmath1 for each redshift slice , and fig . [ f : prelim ] shows a plot of @xmath188 for each slice . we also calculate @xmath199 , the logarithmic slope of the number counts of quasars in terms of flux , by creating a histogram of number counts in terms of the psf magnitude in the @xmath200-band around @xmath201 and calculating the _ actual _ slope around this value . this value for each redshift slice is also listed in table [ t : prelimstat ] . the expression for @xmath175 in eq . [ e : reddistbasic ] requires the true probability distribution @xmath183 . however , since @xmath202 is small , the second term is subdominant to the first and we can substitute @xmath183 with @xmath195 in the second term , giving us @xmath203\pi_{i,\rm prelim}(z')dz'\ , .\end{aligned}\ ] ] @xmath183 can not be similarly substituted for in the first term , so we must estimate @xmath204 using lss data . we estimate @xmath205 as nearly constant , writing @xmath206 where @xmath207 is a constant . generally , @xmath207 would be a piecewise function of @xmath127 with several breaks in order to estimate @xmath207 precisely ; however , we were able to estimate @xmath207 as a constant function with very small uncertainties . we estimate @xmath207 in each redshift slice by constraining its effect on the quasar clustering . we begin by estimating the quasar correlation function @xmath208 in each redshift slice using the method presented in @xcite . specifically , we use the estimator @xmath209 along with its variance given in @xcite , given by @xmath210 where @xmath211 , @xmath212 , and @xmath213 are properly normalized histograms of the number of data - data pairs , data - random pairs , and random - random pairs , respectively , binned in terms of angular separation . we use the following expression for the poisson uncertainty in our estimator @xmath214 = \frac{2}{n_r(n_r-1)rr}+\frac{2}{n(n-1)dd}\ , , \end{aligned}\ ] ] where @xmath27 ( @xmath215 ) is the number of data ( random ) points used to calculate @xmath211 ( @xmath213 ) . note that for the stochastic calculations we use 25 simulations with the number of random points equal to twice the number of data points . we calculate @xmath208 in 10 logarithmic bins of equal ( logarithmic ) size in the range @xmath216 . we avoid estimating @xmath208 at smaller angles due to nonlinearities and at larger angles due to possible non - gaussianity effects on large scales . we compare this estimate of @xmath208 to the expression @xmath217 ^ 2\ , , \end{aligned}\ ] ] where we substitute eq . [ e : amp ] for the first term in eq . [ e : reddistapprox ] to calculate @xmath218 in terms of @xmath219 . we include redshift - space distortions in @xmath218 . for calculating @xmath220 we use the limber approximation for @xmath221 mpc@xmath68 mpc@xmath68 the summand vanishes for @xmath222 , which removes all terms over which the limber approximation is not valid . ] by converting the sum to an integral , replacing @xmath223 , and replacing @xmath224 . we can write @xmath208 as a sum of terms linear and quadratic in @xmath207 , which allows us to fit for this parameter for each redshift slice . we also add to this model a constant term @xmath225 due to systematics effects for which we also fit for redshift slice z03 because it caused @xmath207 to not be constrainable . ] . in table [ t : ai ] , we list estimates for the @xmath207 , from which we derive the redshift distributions @xmath175 shown in fig . [ f : redplot1 ] . .[t : redshifts1 ] properties of the 3 qso photometric redshift slices ; @xmath226 is the photometric redshift range , and @xmath227 is the mean ( true ) redshift of the slice , and @xmath228 is the number of qsos in the redshift slice . [ cols="^,^,^,^,^",options="header " , ] _ top _ : the qso density in the photometric redshift slices z01 , z02 , and z03 , respectively . the 180@xmath35 radius caps are displayed in a lambert azimuthal equal - area projection , with the north galactic pole at the map center , @xmath229 at right , and @xmath230 at bottom . _ bottom _ : the qso density in redshift slice z03 with no kde qso density cut . notice the large - scales fluctuations that would ruin scale - dependent bias measurements.,title="fig:",width=528 ] _ top _ : the qso density in the photometric redshift slices z01 , z02 , and z03 , respectively . the 180@xmath35 radius caps are displayed in a lambert azimuthal equal - area projection , with the north galactic pole at the map center , @xmath229 at right , and @xmath230 at bottom . bottom _ : the qso density in redshift slice z03 with no kde qso density cut . notice the large - scales fluctuations that would ruin scale - dependent bias measurements.,title="fig:",width=153 ] the predicted qso angular power spectra for null and nonzero @xmath231 . the top - left and bottom - right panels are both for the z01 redshift bin , except that the bottom - right panel s @xmath15-range ends at @xmath232 to see the low-@xmath15 behavior of the spectrum . the solid lines are the spectra for the gaussian case , and the dotted ( dashed ) lines represent the @xmath233 case . the dot - dashed line represents @xmath234 . we normalized the @xmath235 cases such that the variances of number density fluctuations within a pixel ( @xmath236 ) for these cases are equal to the value for @xmath84.,title="fig:",scaledwidth=80.0% ] the measured cross - correlation angular power spectra between all three qso redshift slices . the crosses are the measured spectra with 1@xmath1 errors and the solid lines are the predicted spectra . note that the predicted spectra are not exactly zero at smaller scales because the redshift distributions have small overlaps.,scaledwidth=80.0% ] the measured cross - correlation angular power spectra between z01 and z02 as well as z01 and z03 with systematic templates included . the crosses are the measured spectra with 1@xmath1 errors and the solid lines are the predicted spectra . note that the predicted spectra are not exactly zero at smaller scales because the redshift distributions have small overlaps.,scaledwidth=80.0% ]	primordial non - gaussianity of local type is predicted to lead to enhanced halo clustering on very large scales . photometric quasars , which can be seen from cosmological redshifts @xmath0 even in wide - shallow optical surveys , are promising tracers for constraining non - gaussianity using this effect . however , large - scale systematics can also mimic this signature of non - gaussianity . in order to assess the contribution of systematic effects , we cross - correlate overdensity maps of photometric quasars from the _ sloan digital sky survey _ ( sdss ) data release 6 ( dr6 ) in different redshift ranges . we find that the maps are significantly correlated on large scales , even though we expect the angular distributions of quasars at different redshifts to be uncorrelated . this implies that the quasar maps are contaminated with systematic errors . we investigate the use of external templates that provide information on the spatial dependence of potential systematic errors to reduce the level of spurious clustering in the quasar data . we find that templates associated with stellar density , the stellar color locus , airmass , and seeing are major contaminants of the quasar maps , with seeing having the largest effect . using template projection , we are able to decrease the significance of the cross - correlation measurement on the largest scales from 9.2@xmath1 to 5.4@xmath1 . although this is an improvement , the remaining cross - correlation suggests the contamination in this quasar sample is too great to allow a competitive constraint on @xmath2 by correlations internal to this sample . the sdss quasar catalog exhibits spurious number density fluctuations of @xmath32% rms , and we need a contamination level less than 1% ( 0.6% ) in order to measure values of @xmath4 less than 100 ( 10 ) . properly dealing with these systematics will be paramount for future large scale structure surveys that seek to constrain non - gaussianity .
there has been much progress in recent years in model - independent calculations of heavy meson decays . by using heavy quark effective theory ( hqet ) @xcite and lattice gauge theory , we can now make some definite predictions for certain processes , in a limited kinematic range . knowledge of these decays are extremely important for particle physics , not just in their own right , but in measuring fundamental parameters such as @xmath4 and @xmath5 . unfortunately , at present time , it is not always possible to use these techniques for all kinematic situations . one example are the isgur - wise ( iw ) functions , which relate all the different form factors for heavy - to - heavy decays to a single function , at leading order in @xmath6 . these functions can , in principle , be determined by lattice computations but , because of limited computing resources , they have only been computed for decays into the ground state and the error is still sizable . for the moment hqet can be used to produce model independent results at zero recoil but , away from this point , the iw functions are unknown . therefore , for the time being , we must rely on models . we present here a study of @xmath7 decaying into excited @xmath8s and a determination of the corresponding iw functions , following the work of ref . our study is based on the quark model proposed in ref . @xcite , where a dirac equation was used to describe the light quark in the potential of the heavy quark and determine masses and wave functions of excited mesons . we use these wavefunctions to calculate the leading order iw functions for @xmath9 decays , where @xmath1 is a spin 0 2 charmed meson or its first radially excited state . the iw functions are computed explicitly by a three dimensional numerical integration of the relevant matrix elements expressed in terms of the wave functions derived in ref . @xcite . the paper is organized as follows . in section ii we set up the formalism and in section iii we discuss the actual calculation of the iw functions . in section iv we discuss the results and compare with the literature . finally in section v we conclude . due to heavy quark symmetry , we know a great deal about the spectroscopy of mesons containing a heavy quark @xmath10 . at leading order in @xmath11 , the spin and parity of the light degrees of freedom , @xmath12 , are good quantum numbers . thus the particles appear in multiplets labeled by both the total spin @xmath13 and @xmath12 . in this limit , the mesons come in degenerate doublets with total spin @xmath14 , where @xmath15 is the heavy quark spin , a conserved quantum number at leading order in hqet . in table [ particles ] we list the low spin charmed mesons considered in this paper , their masses and corresponding quantum numbers in hqet . in those cases where the experimental mass is not known we report , in brackets , the mass predicted in ref . @xcite . for the mass of the decaying @xmath7 meson we adopt the pdg value @xmath16 gev . .charmed meson spin multiplets ( @xmath17 ) . the masses are experimental when possible , otherwise they are calculated from the model . the primed masses are for the first radial excited states . the masses in parenthesis are predictions of the in the model . [ cols="^,^,^,^,^,^,^,^",options="header " , ] for the radial excitations of the other doublets @xmath18 we do not have any normalization requirements at zero recoil ( since the matrix elements vanish at leading order because of the heavy quark spin symmetry ) . in fig . [ d2all ] we show the iw functions for @xmath19 , @xmath20 , @xmath21 and @xmath22 respectively . information on the decays of @xmath3 mesons into charmed mesons is important for a number of reasons . using heavy quark effective theory , these decays can be written in terms of isgur - wise functions which parameterized the form factors . these functions can not be calculated except , eventually , on the lattice . in this paper we derived the isgur - wise functions of @xmath0 within a quark model , where @xmath1 can be any spin 0 2 charmed meson or one of their first radially excited states . our results are compared with independent predictions found in the literature . our model differs from the others because we derived the wavefunctions of the @xmath3 meson and excited charmed mesons in a relativistic fashion by fitting the experimental spectrum with model predictions . our results for the isgur - wise function , @xmath23 , for @xmath3 to @xmath24 decays is consistent with other model predictions but its slope is milder . our result is also consistent with the preliminary lattice results of ref . @xcite . for decays into the @xmath25-waves the situation is even more uncertain since different models predict a wide varieties of results . we believe that further model independent investigations are required in order to put more constraints important phenomenological quantities . we want to thank estia eichten for useful discussions . fermilab is operated by ura under doe contract no . de - ac02 - 76ch03000 . in this appendix we collect the relations between the matrix elements of the form @xmath26 to expressions in terms of the isgur - wise functions in hqet at leading order in @xmath27 and @xmath28 , in analogy with eq . ( [ matrix1 ] ) . the matrices containing the states for the different doublets are shown in eqs . ( [ hminus ] ) and ( [ hplus]-[fpm ] ) . the choices of dirac structure are @xmath29 . in this appendix @xmath30 is the polarization vector for spin 1 particles and @xmath31 is the polarization tensor for spin 2 particles , while @xmath32 is the usual minkowskian antisymmetric tensor . @xmath33 is the velocity of the decaying @xmath3 meson and @xmath34 is the velocity of the charmed decay product . all matrix elements not explicitly shown are zero : @xmath35 \xi(w ) \\ \nonumber \\ \langle\ ; d_0({\raisebox{0.4ex}{\textrm{\footnotesize $ 1$}}/ \raisebox{-0.4ex}{\textrm{\footnotesize $ 2$}}}^{+})\;\|\;\bar c \gamma^5 b \;\|\ ; \bar b \;\rangle & = & 2 ( w-1 ) \tau_{1/2}(w ) \\ \langle\ ; d_0({\raisebox{0.4ex}{\textrm{\footnotesize $ 1$}}/ \raisebox{-0.4ex}{\textrm{\footnotesize $ 2$}}}^{+})\;\|\;\bar c \gamma^\mu \gamma^5 b \;\|\ ; \bar b \;\rangle & = & 2 ( v - v')^\mu \tau_{1/2}(w ) \\ \nonumber \\ \langle\ ; d_1^\ast({\raisebox{0.4ex}{\textrm{\footnotesize $ 1$}}/ \raisebox{-0.4ex}{\textrm{\footnotesize $ 2$}}}^{+})\;\|\;\bar c b \;\|\ ; \bar b \;\rangle & = & - 2 \epsilon^\nu v_\nu \tau_{1/2}(w ) \\ \langle\ ; d_1^\ast({\raisebox{0.4ex}{\textrm{\footnotesize $ 1$}}/ \raisebox{-0.4ex}{\textrm{\footnotesize $ 2$}}}^{+})\;\|\;\bar c \gamma^\mu b \;\|\ ; \bar b \;\rangle & = & 2 [ \epsilon^\mu ( w-1)-\epsilon^\nu v_\nu v'^\mu ] \tau_{1/2}(w ) \\ \langle\ ; d_1^\ast({\raisebox{0.4ex}{\textrm{\footnotesize $ 1$}}/ \raisebox{-0.4ex}{\textrm{\footnotesize $ 2$}}}^{+})\;\|\;\bar c \gamma^\mu \gamma^5 b \;\|\ ; \bar b \;\rangle & = & 2 i \epsilon^{\mu\alpha\beta\gamma}\epsilon_\alpha v_\beta v'_\gamma \tau_{1/2}(w ) \\ \nonumber \\ \langle\ ; d_1({\raisebox{0.4ex}{\textrm{\footnotesize $ 3$}}/ \raisebox{-0.4ex}{\textrm{\footnotesize $ 2$}}}^{+})\;\|\;\bar c b \;\|\ ; \bar b \;\rangle & = & -\sqrt{2 } \epsilon^\nu v_\nu ( 1+w ) \tau_{3/2}(w ) \\ \langle\ ; d_1({\raisebox{0.4ex}{\textrm{\footnotesize $ 3$}}/ \raisebox{-0.4ex}{\textrm{\footnotesize $ 2$}}}^{+})\;\|\;\bar c \gamma^\mu b \;\|\ ; \bar b \;\rangle & = & \sqrt{\frac12 } [ \epsilon^\mu ( 1-w^2)-3\epsilon^\nu v _ \nu v^\mu + ( w-2 ) \epsilon^\nu v _ \nu v'^\mu\ \tau_{3/2}(w ) \\ \langle\ ; d_1({\raisebox{0.4ex}{\textrm{\footnotesize $ 3$}}/ \raisebox{-0.4ex}{\textrm{\footnotesize $ 2$}}}^{+})\;\|\;\bar c \gamma^\mu \gamma^5 b \;\|\ ; \bar b \;\rangle & = & -\frac{i}{\sqrt{2}}(1+w)\epsilon^{\mu\alpha\beta\gamma } \epsilon_\alpha v_\beta v'_\gamma \tau_{3/2}(w ) \\ \nonumber \\ \langle\ ; d_2^\ast({\raisebox{0.4ex}{\textrm{\footnotesize $ 3$}}/ \raisebox{-0.4ex}{\textrm{\footnotesize $ 2$}}}^{+})\;\|\;\bar c \gamma^5 b \;\|\ ; \bar b \;\rangle & = & \sqrt{3 } \epsilon^{\alpha\beta}v_\alpha v_\beta \tau_{3/2}(w ) \\ \langle\ ; d_2^\ast({\raisebox{0.4ex}{\textrm{\footnotesize $ 3$}}/ \raisebox{-0.4ex}{\textrm{\footnotesize $ 2$}}}^{+})\;\|\;\bar c \gamma^\mu b \;\|\ ; \bar b \;\rangle & = & -i \sqrt{3 } \epsilon^{\mu\alpha\beta\gamma } \epsilon_{\alpha\sigma}v^\sigma v_\beta v'_\gamma \tau_{3/2}(w ) \\ \langle\ ; d_2^\ast({\raisebox{0.4ex}{\textrm{\footnotesize $ 3$}}/ \raisebox{-0.4ex}{\textrm{\footnotesize $ 2$}}}^{+})\;\|\;\bar c \gamma^\mu \gamma^5 b \;\|\ ; \bar b \;\rangle & = & v_\alpha [ \epsilon^{\alpha \beta } v_\beta v'^\mu -\sqrt{3 } ( 1+w ) \epsilon^{\alpha\mu } ] \tau_{3/2}(w ) \\ \nonumber \\ \langle\ ; d_1({\raisebox{0.4ex}{\textrm{\footnotesize $ 3$}}/ \raisebox{-0.4ex}{\textrm{\footnotesize $ 2$}}}^{-})\;\|\;\bar c \gamma^5 b \;\|\ ; \bar b \;\rangle & = & \sqrt{\frac23 } ( 1-w)\epsilon^\nu v_\nu \kappa(w ) \\ \langle\ ; d_1({\raisebox{0.4ex}{\textrm{\footnotesize $ 3$}}/ \raisebox{-0.4ex}{\textrm{\footnotesize $ 2$}}}^{-})\;\|\;\bar c \gamma^\mu b \;\|\ ; \bar b \;\rangle & = & \frac{i}{\sqrt{6}}(1-w)\epsilon^{\mu\alpha\beta\gamma } \epsilon_\alpha v_\beta v'_\gamma \kappa(w ) \\ \langle\ ; d_1({\raisebox{0.4ex}{\textrm{\footnotesize $ 3$}}/ \raisebox{-0.4ex}{\textrm{\footnotesize $ 2$}}}^{-})\;\|\;\bar c \gamma^\mu \gamma^5 b \;\|\ ; \bar b \;\rangle & = & \sqrt{\frac16 } \epsilon^\nu[(2+w)v_\nu v'^\mu - 3 v_\nu v^\mu+ ( 1-w^2)g_\nu^\mu ] \kappa(w ) \\ \nonumber \\ \langle\ ; d_2^\ast({\raisebox{0.4ex}{\textrm{\footnotesize $ 3$}}/ \raisebox{-0.4ex}{\textrm{\footnotesize $ 2$}}}^{-})\;\|\;\bar c b \;\|\ ; \bar b \;\rangle & = & v_\alpha v_\beta \epsilon^{\alpha\beta } \kappa(w ) \\ \langle\ ; d_2^\ast({\raisebox{0.4ex}{\textrm{\footnotesize $ 3$}}/ \raisebox{-0.4ex}{\textrm{\footnotesize $ 2$}}}^{-})\;\|\;\bar c \gamma^\mu b \;\|\ ; \bar b \;\rangle & = & [ ( 1-w ) \epsilon^{\alpha\mu } v_\alpha + \epsilon^{\alpha\beta } v'^\mu v_\alpha v_\beta ] \kappa(w ) \\ \langle\ ; d_2^\ast({\raisebox{0.4ex}{\textrm{\footnotesize $ 3$}}/ \raisebox{-0.4ex}{\textrm{\footnotesize $ 2$}}}^{-})\;\|\;\bar c \gamma^\mu \gamma^5 b \;\|\ ; \bar b \;\rangle & = & i \epsilon^{\mu\alpha\beta\gamma}\epsilon_{\alpha\rho } v^\rho v_\beta v'_\gamma \kappa(w)\end{aligned}\ ] ] a. f. falk , h. georgi , b. grinstein and m. b. wise , nucl . b * 343 * , 1 ( 1990 ) . j. d. bjorken , slac - pub-5389 _ lectures given at 18th annual slac summer inst . on particle physics , stanford , ca , jul 16 - 27 , 1990_. a. f. falk , nucl . b * 378 * , 79 ( 1992 ) . c. g. boyd , b. grinstein and r. f. lebed , phys . b * 353 * , 306 ( 1995 ) . b * 461 * , 493 ( 1996 ) . d * 56 * , 6895 ( 1997 ) . i. caprini , l. lellouch and m. neubert , nucl . b * 530 * , 153 ( 1998 ) . n. uraltsev , phys . b * 501 * , 86 ( 2001 ) . n. isgur , d. scora , b. grinstein and m. b. wise , phys . d * 39 * , 799 ( 1989 ) . m. wirbel , b. stech and m. wirbel , z. phys . * c29 * , 637 ( 1985 ) ; m. bauer , b. stech and m. wirbel , z. phys . * c34 * , 103 ( 1987 ) ; j.g . korner and g.a . schuler , z. phys . * c38 * , 511 ( 1988 ) , erratum - ibid . * c41 * , 690 ( 1989 ) . a. le yaouanc , l. oliver , o. pene , j. c. raynal and v. morenas , phys . b * 520 * , 25 ( 2001 ) ; phys . b * 520 * , 59 ( 2001 ) . a. k. leibovich , z. ligeti , i. w. stewart and m. b. wise , phys . lett . * 78 * , 3995 ( 1997 ) ; phys . d * 57 * , 308 ( 1998 ) . d. ebert , r. n. faustov and v. o. galkin , phys . b * 434 * , 365 ( 1998 ) . a. deandrea , n. di bartolomeo , r. gatto , g. nardulli and a. d. polosa , phys . rev . d * 58 * , 034004 ( 1998 ) . a. wambach , nucl . phys . b * 434 * , 647 ( 1995 ) . p. colangelo , f. de fazio and n. paver , phys . d * 58 * , 116005 ( 1998 ) . v. morenas , a. le yaouanc , l. oliver , o. pene and j. c. raynal , phys . d * 56 * , 5668 ( 1997 ) . s. godfrey and n. isgur , phys . d * 32 * ( 1985 ) 189 . p. colangelo , g. nardulli and m. pietroni , phys . d * 43 * , 3002 ( 1991 ) . s. anderson _ [ cleo collaboration ] , nucl . phys . a * 663 * , 647 ( 2000 ) . d. abbaneo _ et al . _ [ aleph collaboration ] , arxiv : hep - ex/0112028 , and references therein . n. isgur and m. b. wise , phys . lett . * 66 * , 1130 ( 1991 ) . s. veseli and m. g. olsson , phys . b * 367 * , 302 ( 1996 ) ; phys . d * 54 * , 886 ( 1996 ) . n. isgur and m. b. wise , phys . d * 43 * , 819 ( 1991 ) . d. ebert , r. n. faustov and v. o. galkin , phys . d * 62 * , 014032 ( 2000 ) .	in this paper we compute , within in the context of a relativistic quark model , the isgur - wise functions for exclusive semileptonic @xmath0 decays , where @xmath1 is any charmed mesons with total spin @xmath2 or one of their first excited states . the relevant matrix elements are computed by a direct numerical integration , in coordinate space , of the convolution of the wave function of the @xmath3 meson at rest and the wave function of the @xmath1 meson , boosted according with its recoil factor . our results are compared with other predictions found in the existing literature .
though the first study of electromagnetic wave propagation in periodic structures , written by lord rayleigh , was published already in 1887 , not until very recently have photonic composites suddenly raised an extremely keen interest . the research in this field was triggered by the studies by yablonovitch and john @xcite , published in 1987 and anticipating the existence of complete energy gaps in electromagnetic wave spectra in three - dimensional periodic composites , henceforth referred to as _ photonic crystals_. these can be used for fabricating new optoelectronic devices with photons acting as transport medium . the so - called _ left - handed materials _ , an example of periodic structures characterized by negative effective refractivity @xcite , demonstrate how much the properties of this kind of structure can differ from those of homogeneous materials . another type of periodic composites are structures composed of materials with different elastic properties ; showing an energy gap in their elastic wave spectrum , such composites are referred to as _ phononic crystals _ @xcite . recently , attention has been focused on the search of photonic and phononic crystals in which both the position and the width of the energy gap could be controlled by external factors , such as applied voltage or _ magnetic field_. attempts are made to create photonic crystals in which one of the component materials would be a magnetic @xcite . a magnetic periodic composite , conceived as the magnetic counterpart of a photonic crystal , consists of at least two magnetic materials , with magnon acting as information carrier , therefore such periodic magnetic composites can be referred to as _ magnonic crystals _ ( mc ) . studies of 2d magnonic crystals have already been reported @xcite with scattering centers in the form of `` infinitely '' long cylinders disposed in square lattice nodes ( cylinder and matrix materials being two different ferromagnetics ) and spin wave spectra showing the anticipated gaps . further research was focused on magnetic multilayer systems , which can be regarded as 1d magnonic crystals @xcite . in this paper , we present numerically calculated - and , to our best knowledge , not yet reported in the literature - band structures of _ three - dimensional _ magnonic crystals . due to the complexity of the problem , only the simplest model of 3d magnonic crystal is considered here , represented by a system of ferromagnetic spheres ( which act as scattering centers ) disposed in the nodes of a cubic _ bcc _ crystal lattice and embedded in a different ferromagnetic material ( matrix ) . both the exchange and dipolar interactions are taken into account in our calculations , which are based on the plane wave method and use the linear approximation . as a conclusion , we propose a new magnonic interpretation of experimental results obtained through neutron scattering on spin waves in doped manganites @xcite . , highlighted ) along which the magnonic band spectra are calculated . [ fig1],width=226 ] let s consider an ideal periodic structure consisting of spheres of ferromagnetic * a * embedded in a matrix of ferromagnetic * b. the spheres are assumed to form a 3d periodical lattice of _ bcc _ ( fig . 1a ) type . a static magnetic field , @xmath1 , is applied to the composite along the @xmath2 axis , and assumed to be strong enough to saturate the magnetization of both materials . the lattice constant is denoted by @xmath3 ; the filling fraction , @xmath4 , is defined as the volume proportion of material a * in a unit cell . ferromagnetics * a * and * b * are characterized by two material parameters : the spontaneous magnetization ( @xmath5 and @xmath6 ) , and the exchange constant ( @xmath7 and @xmath8 ) ; both these parameters depend on the position vector @xmath9 : @xmath10 the value of function @xmath11 being 1 inside a sphere , and 0 beyond . in the classical approximation , spin waves are described by the landau - lifshitz ( ll ) equation , taking the following form in the case of magnetic composites ( with damping neglected ) : @xmath12 where magnetization @xmath13 is a function of position vector @xmath14 and time @xmath15 ; @xmath16 stands for the effective magnetic field @xcite acting on magnetization @xmath13 : @xmath17 @xmath18 is the unit vector along the @xmath2 axis ; @xmath19 is the dynamic magnetic field resulting from dipolar interactions ; the third component represents the exchange field . the magnetization vector can be represented as the sum of its static and dynamic components ; the former , @xmath20 , is parallel to the applied field ; the latter , @xmath21 , @xmath22 denoting the precession circular frequency , lies in the plane ( @xmath23 ) : @xmath24 the dynamic dipolar field , @xmath25 , must satisfy the magnetostatic maxwell equations : @xmath26 in magnonic crystals , the position - dependent coefficients in ( [ eq4 ] ) , _ i.e. _ @xmath27 and @xmath28 , are periodic functions of the position vector , which allows us to use in the procedure of solving the ll equation ( [ eq3 ] ) the plane wave method , described in detail in our earlier papers @xcite ( dealing with _ two - dimensional _ magnonic crystals ) . following this scheme , we proceed to fourier - expanding all the periodic functions of the position vector , _ i.e. _ the spontaneous magnetization , @xmath27 , and parameter @xmath29 defined as follows : @xmath30 the dynamic components of the magnetization can be expressed as the product of the periodic envelope function and the bloch factor , @xmath31 ) ( @xmath32 denoting a 3d wave vector ) ; the envelope function can be transformed into the reciprocal space as well . including all the expansions into ( [ eq3 ] ) and ( [ eq6 ] ) leads to the following infinite system of linear equations for fourier coefficients of the dynamic magnetization components , @xmath33 and @xmath34 : @xmath35 m_{s}(\vec{g}-\vec{g } '' ) q(\vec{g } '' - \vec{g } ' ) m_{y \vec{k } } ( \vec{g}'),\nonumber \\ i \omega m_{y \vec{k}}(\vec{g})= -m_{x \vec{k } } ( \vec{g})-\sum_{\vec{g } ' } \frac{(k_{y } + g'_{y})(k_{x}+g'_{x } ) m_{y \vec{k}}(\vec{g}')+(k_{x } + g'_{x})^{2}m_{x \vec{k}}(\vec{g}')}{h_{0}\| \vec{k } + \vec{g } ' \|^{2 } } m_{s}(\vec{g } -\vec{g } ' ) \nonumber \\ - \sum_{\vec{g } ' } \sum_{\vec{g } '' } [ ( \vec{k } + \vec{g } ' ) \cdot ( \vec{k}+\vec{g } '' ) - ( \vec{g}-\vec{g } '' ) \cdot ( \vec{g } - \vec{g } ' ) ] m_{s}(\vec{g}-\vec{g } '' ) q(\vec{g } '' - \vec{g } ' ) m_{x \vec{k } } ( \vec{g } ' ) ; \nonumber \\ \label{eq8}\end{aligned}\ ] ] @xmath36 and @xmath37 denoting the cartesian components of the wave vector @xmath32 and a reciprocal lattice vector @xmath38 , respectively ; a new quantity introduced in ( [ eq8 ] ) is @xmath39 , henceforth referred to as _ reduced frequency _ : @xmath40 the fourier coefficients of spontaneous magnetization @xmath27 and parameter @xmath29 are calculated from the inverse fourier transformation ; in the case of spheres , the resulting formulae for @xmath27 read as follows : @xmath41}{(gr)^{3}}},\\ \mbox { \hspace{5cm}for $ \;\vec{g } \neq 0$}% \end{array } \right.% \label{aa1}\end{aligned}\ ] ] and similarly for @xmath29 ; @xmath42 is the sphere radius ( fig . [ fig1]a ) . obviously , the numerical calculations performed on the basis of ( [ eq8 ] ) involve a finite number of reciprocal lattice vectors @xmath38 in the fourier expansions ; however , we have made sure that the number used is large enough to guarantee good convergence of the numerical results . as indicated by an analysis performed , a satisfactory convergence is obtained already with 343 reciprocal lattice vectors used . ) for a hypothetic _ periodic _ arrangement of ferromagnetic droplets embedded in different ferromagnetic medium ; the sphere - shaped droplets are assumed to form a regular crystal lattice of _ bcc _ type . ( a ) magnonic branches plotted versus the lattice constant , @xmath3 , of the droplet _ bcc _ structure ; the filling fraction is fixed at @xmath43 . a frequency gap ( the shaded region ) is found to exist , and move down with increasing @xmath3 . the superimposed empty rectangle represents the gap revealed experimentally in doped manganites @xcite . ( b ) the magnonic energy branches ( in the same _ crystal _ ) plotted _ versus _ the filling fraction , with the lattice constant fixed at @xmath44.,width=302 ] an energy gap in spin wave spectra has recently been found experimentally in doped manganites @xmath0 and @xmath45 at low ( @xmath46 or @xmath47 ) ion concentrations ( @xmath48 ) @xcite . the appearance of this gap was suggested to be due to doping , which results in a two - phase magnetic structure , with ferromagnetic ( f ) droplets ( associated with the @xmath49 ions ) embedded in a canted antiferromagnetic ( caf ) medium ( showing a non - zero magnetic moment along the direction of the droplet magnetic moments ) . hence , the two _ separate _ energy branches found in the neutron scattering experiments are interpreted as corresponding to two _ independent _ spin - wave excitations , each propagating in one of the two phases . in this paper , we propose an alternative _ magnonic _ explanation of the spin - wave spectrum energy gap found in these experiments . we anticipate that the dispersion branches obtained in the experiment constitute the bottom part of the _ collective _ spin - wave spectrum , involving spin waves propagating _ concurrently through both phases _ of a magnetic composite , resulting from a _ regular _ arrangement of ferromagnetic droplets ( material * a * ) in a matrix with different magnetic properties ( material * b * ) . then , the existence of a spin - wave spectrum gap becomes a natural consequence of the structure periodicity , as in the case of photonic or phononic crystals . below we present results of our calculations ( based on the method discussed in the preceding paragraph ) of the spin - wave spectrum in a _ droplet magnonic crystal _ , the necessary material parameter values being as in the experimental studies @xcite . the results are compared to the spectrum obtained experimentally in @xmath50 . let s assume that the spherical ferromagnetic droplets form a _ bcc _ lattice ( fig . [ fig1]a ) , which we believe to be a good approximation of the actual droplet structure . the spin - wave branches will be computed along the following path in the irreducible part of the first brillouin zone : @xmath51 ( see fig . [ fig1]b ) . the necessary material parameter values , _ i.e. _ those of the spontaneous magnetization and the exchange constant , in both magnetic phases of our model magnonic crystal are as in the experimental studies referred to above . the paper by hennion _ @xcite provides information on the difference , @xmath52 , between the values of magnetization along vector @xmath53 in the droplets and in the matrix : @xmath54 . assuming that the magnetic moment in droplets comes mainly from @xmath49 ions with spin @xmath55 , we get droplet spontaneous magnetization value @xmath56 ; consequently , the matrix magnetization ( @xmath5 less 0.7@xmath57 ) is @xmath58 . the following lattice constant values in @xmath50 are assumed : @xmath59 , @xmath60 , and @xmath61 @xcite . the exchange constant values in droplets and in the matrix , @xmath7 and @xmath8 , respectively , are determined on the basis of the exchange integral , @xmath62 , and the stiffness constant , @xmath63 , as estimated in the study by biotteau _ _ @xcite ; the resulting droplet exchange constant value is @xmath64=0.09 \cdot 10^{-11 } \ ; jm^{-1}$ ] , @xmath65 being the lattice constant recalculated for the manganite pseudocubic elementary cell . the matrix exchange constant value , @xmath8 , is determined by fitting the weighted mean value of the droplet and matrix stiffness constants to the stiffness constant value estimated in @xcite . we assume @xmath66 , which , at filling fraction @xmath67 , corresponds to mean value @xmath68 , close to the value estimated by fitting to the experimental results at temperature @xmath69k . the assumed average value of the isotropic internal field , @xmath70 , lies between the values found by savosta _ @xcite ( 0.064 t ) and yates _ et al . _ @xcite ( 0.35 t ) . plane with two sets of lattice constant @xmath3 and filling fraction @xmath71 values in a magnonic crystal : one , represented by the _ solid _ line , corresponds to an energy gap whose width is exactly @xmath72 ; the other , represented by the _ dotted _ line , corresponds to a gap centered around @xmath73 . the cross - section points indicate the @xmath3 and @xmath71 values at which the computed gap fits the experimental one in terms of both width ( @xmath74 ) and position ( @xmath75 ) . ( b ) magnonic bands calculated on the basis of our model across the brillouin zone , the assumed lattice constant and filling fraction values being @xmath76 , and @xmath77 , respectively . the gap which is found to exist has position and width exactly fitting the experimental data reported in @xcite.,width=302 ] figure [ fig2]a shows the computed magnonic band structure plotted against the droplet lattice constant . the gap center ( @xmath75 ) is found to descend , and the gap itself ( the dark gray region ) to narrow down as the lattice constant increases . the superimposed rectangle represents the experimental gap reported in @xcite ) . the _ gap center _ found experimentally corresponds to our results at @xmath78 ; however , the corresponding gap width resulting from our computations is much above the experimental result , which implies the necessity of fitting the filling fraction as well . [ fig2]b shows the magnonic band structure plotted against the filling fraction @xmath71 ( the lattice constant being fixed at @xmath79 ) . the gap is found to open at @xmath80 to reach its maximum width at @xmath81 and vanish at @xmath82 , which means that one gap width value corresponds to two different filling fraction values . note also that the gap center moves up as the filling fraction increases . hence , the @xmath3 and @xmath71 values corresponding to an energy gap that fits the experimental results will be determined in two steps ( see fig . [ fig3]a ) : first , we shall find the set of @xmath3 and @xmath71 values corresponding to the experimental gap width , @xmath83 ( the solid line in fig [ fig3]a ) ; next , another line ( the dotted one in the figure ) is plotted , corresponding to the gap centered at the experimental value , @xmath73 . the cross section of these two lines indicates the @xmath3 and @xmath71 values at which both @xmath75 and @xmath74 are fitted . as the lines cross at two points : ( @xmath76 , @xmath77 ) and ( @xmath84 , @xmath85 ) , two solutions are obtained , one corresponding to the droplet radius @xmath86 , the other to @xmath87 . the former solution proves close beyond expectation to the experimental result reported in @xcite ( fig . 9 ) , where @xmath50 droplet radius is estimated at @xmath88 . the corresponding mean droplet spacing estimated in @xcite is @xmath89 , also quite close to our result , @xmath76 . therefore , the complete set of structural and material parameters characterizing our droplet magnonic model of @xmath50 is : @xmath56 , @xmath58 , @xmath90 , @xmath91 , @xmath70 , @xmath92 , @xmath93 . the corresponding full spin - wave spectrum is shown in fig . an energy gap , existing indeed between the first band and the second one , corresponds very well , in terms of both position and width , to that found experimentally @xcite . this result allows to propose a working hypothesis that low - doped manganites could be regarded as magnonic crystals existing in nature . obviously , further research is required for verification of this hypothesis ; therefore , in another paper , we are going to investigate the effect of non - spherical droplet shape on the magnonic spectrum , and to consider a real canted antiferromagnetic matrix , rather than the effective ferromagnetic one considered here . however , already the results obtained here on the basis of the approximate model are extremely promising . 99 e. yablonovitch , phys . * 58 * , 2059 ( 1987 ) . s. john , phys . lett . * 58 * , 2486 ( 1987 ) . j. b. pendry , d. r. smith , physics today * 57 * , 37 ( 2004 ) . m. s. kushwaha , p. halevi , l. dobrzynski , b. djafari - rouhani , phys . lett . * 71 * , 2022 ( 1993 ) . a. saib , d. vanhoenacker - janvier , i. huynen , a. encinas , l. piraux , e. ferain , r. legras , appl . phys . lett . * 83 * , 2378 ( 2003 ) . nikitov , ph . tailhades , optics communications * 199 * , 389 ( 2001 ) . landau and e.m . lifshitz , phys . z. sowjetunion * 8 * , 153 ( 1935 ) . j. o. vasseur , l. dobrzynski , b. djafari - rouhani , h. puszkarski , phys . b * 54 * , 1043 ( 1996 ) . m. krawczyk , h. puszkarski , acta physica polonica a * 93 * , 805 ( 1998 ) . h. puszkarski , m. krawczyk , solid state phenomena * 94 * , 125 ( 2003 ) . gulyaev , s.a . nikitov , l.v . zhivotovskii , a.a klimov , p. tailhades , l. presmanes , c. bonningue , c.s . tsai , s.l . vysotskii , y.a.filimonov , j.e.t.p . letters * 77 * , 670 ( 2003 ) . m. shamonin , a. snarskii and m. zhenirovskiyy , ndt&e international * 37 * , 35 ( 2004 ) . h. puszkarski , m. krawczyk , physics letters a * 282 * , 106 ( 2001 ) . kruglyak , a.n . kuchko , j. magn . magn . mater . * 272 - 276 * , 302 ( 2004 ) . m. hennion , f. moussa , g. biotteau , j. rodriguez - carvajal , l. pinsard , a. revcolevschi , phys . lett . * 81 * , 1957 ( 1998 ) . g. biotteau , m. hennion , f. moussa , j. rodriguez - carvajal , l. pinsard , a. revcolevschi , y. m. mukovskii , d. shulyatev , phys . b * 64 * , 104421 ( 2001 ) . kober - lehouelleur , f. moussa , m. hennion , a. ivanov , l. pinsard - gaudart and a. revcolevschi , phys . b * 70 * , 144409 ( 2004 ) . m.m . savosta , p. novak , phys . rev . 87 * , 137204 ( 2001 ) . yates , c. kapusta , p.c . riedi , l. ghivelder , l.f . cohen , j. magn . magn . mater . * 260 * , 105 ( 2003 ) .	a theory of three - dimensional ( 3d ) hypothetical magnonic crystal ( conceived as the magnetic counterpart of the well - known photonic crystal ) is developed and applied to explain the existence of a spin - wave frequency gap recently revealed in low - doped manganites @xmath0 by neutron scattering . a successful confrontation with the experimental results allows us to formulate a working hypothesis that certain manganites could be regarded as 3d magnonic crystals existing in nature .
there is a growing need for a complete census of agn activity . the seminal discovery that every massive galaxy in the local universe harbors a super - massive black hole ( smbh ; @xmath7 @xmath8 ) implies that all galaxies have hosted agn activity at some time during the past @xmath9 13 gyrs ( e.g.,@xcite ; @xcite ) . to accurately trace how and when these smbhs grew requires a detailed census of agn activity that will provide , amongst other things , constraints on the efficiency and duty cycle of smbh growth . the finding that the mass of the smbh is proportional to that of the galaxy spheroid also indicates that the growth of smbhs and their host galaxies are regulated in some way ( e.g. , @xcite ; @xcite ) . whether or not this growth occured concordantly is unknown but comparisons between the rate of star formation and smbh growth , both in individual objects and from a cosmic census , will provide important insight . since obscured agns outnumber unobscured agns by a factor of @xmath9 310 , any `` complete '' census of agn activity must be sensitive to the detection and identification of obscured sources . x - ray emission is relatively impervious to obscuration , at least at high energies , and the detection of luminous or hard x - ray emission provides an almost unambiguous identification of agn activity even in the absence of source redshifts . for this reason , the most complete census of agn activity to date is provided by deep x - ray surveys . for example , the 2 ms _ chandra _ deep field north ( cdf - n ; @xcite ) survey identifies a @xmath10 times larger agn source density ( @xmath9 7200 deg@xmath11 , @xcite ) than found in optical surveys ; see fig . 1a for the sensitivity of the cdf - n to agn objects . however , although deep x - ray surveys have proved stunningly effective at identifying both obscured and unobscured agn activity ( see @xcite for an overall review ) , there is clear evidence that a large fraction of the agn population remains undetected in the ( observed frame ) band probed by these surveys : ( 1 ) about half of the x - ray background is unresolved at @xmath12 kev ( @xcite ) , ( 2 ) the obscured : unobscured agn ratio is much lower than that found for comparably luminous agns in the local universe ( e.g. , @xcite ) , and ( 3 ) few compton - thick agns ( @xmath13 cm@xmath11 ) have been identified , even though they comprise @xmath9 50% of the agn population in the local universe ( e.g. , @xcite ; @xcite ) . many of these x - ray undetected agns are likely to be intrinsically luminous sources that are heavily obscured ( @xmath1510 @xmath16 cm@xmath11 ) by an optically and geometrically thick torus of gas and dust ( e.g. , @xcite ; see fig . 1a for the detectability of the intrinsically luminous agn in ngc 1068 in the observations ) . although the primary emission from these obscured agns will be absorbed by the torus and rendered weak or invisible at x - ray optical wavelengths , it will heat the dust grains and re - emit this emission relatively isotropically in the infrared ( ir ) waveband ( rest - frame @xmath17 @xmath18 m ; e.g. , @xcite ; @xcite ) . the intrinsically luminous but x - ray undetected agns should therefore be detected in deep ir surveys ; see fig . radio - loud agns or to agn activity in nearby galaxies ( but see @xcite ) . future radio observatories such as the square kilometer array will provide an extremely sensitive agn census but will not be _ fully _ operational for @xmath10 years ( e.g. , @xcite ) ; see http://www.skatelescope.org/. ] since star formation can also produce luminous ir emission it is necessary to distinguish between agns and starburst galaxies . a number of recent studies have demonstrated that ir colours can often distinguish between agn and star - formation activity using crude ir data , with the former having `` hotter '' ir spectral energy distributions ( seds ; e.g. , @xcite ; @xcite ) . however , since a large fraction of the agn population in the local universe has ir colors similar to those of starburst galaxies , other indicators ( e.g. , the detection of high - excitation emission lines , hard x - ray emission from stacking analyses of samples of objects , and luminous unresolved radio emission ) are also required to identify agn activity from star formation . the primary aim of this article is to investigate the prospects for selecting x - ray undetected agn with _ observations in the ultra - deep cdf - n / goods - n field to increase the agn census beyond that identified from the x - ray data alone . here we select x - ray undetected agn candidates using the _ spitzer _ goods - n observations and then search for the presence of hard x - ray emission using x - ray stacking techniques . from these results we place constraints on the demographics and properties of agn activity in the ir galaxy population . the _ spitzer _ goods observations ( p.i . m. dickinson ) were performed as part of the _ spitzer _ legacy program and cover @xmath9 160 arcmin@xmath19 of the most sensitive regions in each of the _ chandra _ deep fields ( the cdf - n and the cdf - s ) . the irac 3.68 @xmath18 m and mips 24 @xmath18 m _ spitzer _ observations are the deepest across the sky , achieving typical 1 @xmath20 sensitivities of @xmath9 0.030.3 @xmath18jy ( 3.68 @xmath18 m ) and @xmath9 5 @xmath18jy ( 24 @xmath18 m ) . here we focus on the cdf - n / goods - n field , which has the deepest x - ray coverage to date . to provide maximum sensitivity to x - ray faint agn activity we restrict our analyses to the region within 5.5 arcmin of the _ chandra _ aimpoint ( 95 arcmin@xmath19 ) , where the x - ray sensitivity and angular resolution is optimal ; see figs . 18 & 19 of @xcite . there are 223 x - ray sources , @xmath9 1800 mips 24 @xmath18 m sources , and @xmath9 3400 irac 8 @xmath18 m sources within this region . the x - ray sources were matched to the ir sources on the basis of the x - ray source positional uncertainties given in @xcite . the positional uncertainites were linearly increased to give a 3% spurious source match fraction ; extensive monte - carlo tests were used to determine the spurious match fraction . overall , 78% and 87% of the x - ray sources were matched to 24 @xmath18 m and 8 @xmath18 m sources , respectively . using the multi - wavelength source classification of @xcite we distinguished between x - ray detected agns and x - ray detected starbursts . the x - ray sources comprise 9.4% of the 24 @xmath18 m source population ( 7.0% x - ray agns and 2.4% x - ray starbursts ) and 5.6% of the 8 @xmath18 m source population ( 4.3% x - ray agns and 1.3% x - ray starbursts ) . 2 shows the x - ray vs 24 @xmath18 m properties of the x - ray sources , including the x - ray agns and x - ray starburst fraction for different 24 @xmath18 m flux density bins . as expected , the majority of the x - ray agns have x - ray ir flux ratios higher than those of the x - ray starbursts . . 2 also shows a clear drop in the x - ray detected fraction of ir sources towards the faintest ir fluxes . this is at least partially due to the linear decrease in the observed x - ray ir flux ratio with 24 @xmath18 m flux density , suggesting that the majority of the x - ray undetected agn population will have faint ir fluxes.m flux - density bins but it is at least qualitatively similar to the conclusions drawn from the agn population synthesis model of @xcite . ] evidence for an x - ray undetected agn population can be explored using stacking techniques . the principle behind this approach is to stack the x - ray data of individually undetected _ spitzer _ sources and determine the probability that the stacked x - ray signal could be produced by the background ( e.g. , @xcite ; the background is determined from a monte - carlo analysis of source - excluded regions around the stacked sources . the expected signature of a significant obscured agn population is a detection in the hard ( 28 kev ) band and a comparatively flat x - ray spectral slope ( @xmath21 ) . by comparison , unobscured / low - luminosity agn and typical starburst galaxies have softer x - ray spectral slopes ( @xmath22 2 ; e.g. , @xcite ; @xcite ) . we note that neutron - star x - ray binary populations can also have flat x - ray spectral slopes ( e.g. , @xcite ) , however , it seems unlikely that the integrated x - ray emission from a starburst galaxy will be dominated by this comparatively rare galactic population . for the stacking analyses we have used the procedure outlined in @xcite and stacked _ spitzer _ galaxies in the soft ( 0.52 kev ) , hard ( 28 kev ) , and full ( 0.58 kev ) bands . the background was determined using 1,000 monte - carlo trials for each stacked source and the x - ray counts were measured using apertures based on the 90% encircled - energy radius . stacking all of the x - ray undetected ir sources we obtain highly significant detections ( @xmath23 10 @xmath20 ) in all three x - ray bands , corresponding to an overall x - ray spectral slope of @xmath24 and suggesting that some of the stacked sources are likely to be obscured agns ( contrast with the @xmath25 result obtained by @xcite using shallower 15 @xmath18 m data over a smaller area ) ; this is further confirmed by a 3.5 @xmath20 detection in the ultra - hard ( 48 kev ) band . we also stacked the data in the 68 kev band and obtained a non detection in this band ; the 3 @xmath20 upper limit corresponds to @xmath26 20% of the 68 kev background . if a substantial fraction of the x - ray background is unresolved in the 68 kev band by current x - ray surveys , as suggested by the analyses of @xcite , then the other contributing sources must be either ir - faint ( @xmath27 @xmath18jy ) or comparatively rare in @xmath9 100 arcmin@xmath19 regions . a basic interpretation of the data in fig . 2 is that x - ray undetected agns will be predominantly detected at the faintest ir fluxes . to test this simple prediction we stacked the x - ray data for individually undetected sources in different 24 @xmath18 m flux - density bins ; see fig . the 28 kev stacked flux generally increases towards fainter 24 @xmath18 m flux densities , suggesting an increase in the contribution from obscured agn activity , with the @xmath28 75225 @xmath18jy bin producing the strongest and hardest x - ray signal . however , the softer x - ray spectral slope for sources with @xmath29 2575 @xmath18jy implies that starburst galaxies probably dominate the stacked x - ray emission at the faintest ir fluxes . furthermore , the comparatively weak 28 kev flux suggests that there may be fewer obscured agns than at @xmath28 75225 @xmath18jy , possibly indicating a genuine drop in the number of obscured agns at low ir luminosities or at high redshifts ( @xmath30 75 @xmath18jy corresponds to @xmath31 @xmath3 at @xmath32 ) . these results are qualitatively similar to those expected from the agn population synthesis model of @xcite and the interpretation of shallower _ spitzer _ and x - ray data by @xcite . more detailed analyses are required to provide definitive conclusions . we can also explore the average x - ray properties of the _ spitzer _ sources as a function of ir spectral slope ( @xmath33 , where @xmath34 @xmath33 @xmath35 ) . in the local universe many agns have flat ir spectral slopes while starburst galaxies are often distinguished by steeper ir spectral slopes . @xcite investigated the detailed properties of _ spitzer _ galaxies with flat ir spectral slopes ( @xmath36 ) in the 1 ms cdf - s field and deduced that only @xmath9 50% of these ir - selected agns are detected in the x - ray band . although the flat ir spectral slopes of these sources clearly indicate that they host agn activity , they are comparatively rare ( @xmath9 800 sources deg@xmath11 , including both x - ray detected and x - ray detected agns ) and unlikely to dominate the unresolved @xmath37 kev background . as a comparison , the range of ir spectral slopes for the x - ray detected agn in the cdf - n / goods - n dataset is @xmath381.0 , and only @xmath9 30% have @xmath36 , indicating that many have ir spectral properties more consistent with those of starburst galaxies . since the aim of the study here is to provide the most complete census of agn activity , we have stacked the x - ray data of _ spitzer _ galaxies over a wide range of ir spectral slopes ( @xmath36 , @xmath390.0 , @xmath40 0.00.5 , @xmath40 0.51.0 ) . the x - ray spectral slope and 28 kev flux density of the stacked sources are shown in fig . all are significantly detected in the 28 kev band but the strongest and hardest x - ray signal comes from the sources with the steepest ir spectral slope ( @xmath40 0.51.0 ) . the steep ir spectral slope of these sources does not obviously imply the presence of agn activity but the hard x - ray signal ( @xmath41 1.2 ) indicates that a considerable fraction of the sources host obscured agn activity . motivated by the discovery that sources with @xmath42 75 @xmath18jy show greater evidence for x - ray undetected agn activity ( see fig . 3a ) , we further stacked the sources in two different 24 @xmath18 m flux bins ( @xmath28 2575 @xmath18jy and @xmath30 75 @xmath18jy ) . the flatter x - ray spectral slope ( @xmath43 ) and larger 28 kev flux density of the @xmath42 75 @xmath18jy sources with @xmath40 0.51.0 clearly indicates that the majority of the obscured agns reside in this population . the x - ray stacking analyses provide evidence that a large x - ray undetected agn population does indeed reside in the _ spitzer _ galaxy population . the dominant agn population has an ir spectral slope of @xmath40 0.51.0 and @xmath30 75 @xmath18jy . their ir colors are more similar to starburst galaxies than agns but x - ray stacking analyses indicate that they host obscured agn activity . we can gain further insight into the properties of these sources by investigating their median redshifts , luminosities , and source densities . the median spectroscopic redshift of the @xmath40 0.51.0 _ spitzer _ galaxies is @xmath1 0.8 , which corresponds to an median luminosity of @xmath44 @xmath3 ; the majority of the redshifts were taken from @xcite . assuming this median redshift , the 28 kev luminosity of the stacked sources with @xmath40 0.51.0 and @xmath30 75 @xmath18jy is @xmath45 erg s@xmath6 ( rest - frame 3.614.4 kev ) . this x - ray luminosity should be considered a lower limit since ( 1 ) many of the stacked galaxies will not be agns and their inclusion in the stacking analysis will decrease the average 28 kev flux , and ( 2 ) the luminosity has not been corrected for the presence of instrinsic absorption . these two issues are briefly considered below . we can not directly determine the `` dilution '' to the hard - band x - ray signal from non - agn activity , however , we can place constraints from the non - detection of individual sources in the x - ray data . for example , if the x - ray undetected agn represented just 1% of the stacked ir galaxies but contributed all of the 28 kev flux then they would have individual 28 kev luminosities @xmath9 2 orders of magnitude higher and would be individually detected in the x - ray data . therefore , since the stacked 28 kev flux is an order of magnitude below the individual x - ray sensitivity limit , these x - ray undetected agns are likely to represent @xmath10% of the stacked ir galaxies . this indicates an average _ observed _ ( i.e. , uncorrected for absorption ) 28 kev luminosity of @xmath45@xmath46 erg s@xmath6 and places lower limits on the x - ray undetected obscured agn source density of @xmath9 5005,000 deg@xmath11 ( for agns with @xmath40 0.51.0 and @xmath30 75 @xmath18jy ) . a crude constraint on the amount of absorption towards the agns can be determined from the average x - ray spectral slope for a given agn model ( e.g. , an observed x - ray spectral slope of @xmath41 1.0 from a @xmath1 0.8 agn would imply @xmath47 @xmath48 cm s@xmath11 for a simple intrinsic agn power - law spectrum of @xmath25 ) . since the soft - band flux almost certainly has a large contribution from star - formation activity , this approach effectively provides a lower limit on the true absorbing column density . another approach , albeit very uncertain , is to estimate the _ intrinsic _ ( i.e. , corrected for absorption ) x - ray luminosity of the agns from the ir luminosities and then compare this to the observed x - ray luminosity . under the assumption that 10100% of the ir luminosity is powered by an agn , the median x - ray luminosity of these sources estimated from the @xcite sed is @xmath49@xmath50 erg s@xmath6 , suggesting large amounts of absorption ( @xmath51 cm@xmath11 ) . although these analyses on the properties of the x - ray undetected agns are crude , their estimated properties are qualitatively consistent with those required by @xcite to produce the unresolved x - ray background ( i.e. , @xmath1 0.51.5 , @xmath52@xmath53 cm@xmath11 , @xmath54@xmath50 erg s@xmath6 ) . interestingly , the median redshift , ir luminosity , and x - ray luminosity range of the x - ray detected agns with @xmath40 0.51.0 are also similar to those estimated for the x - ray undetected agns with @xmath40 0.51.0 . this is entirely consistent with that expected for the unified agn model , where the x - ray undetected agns represent the more heavily obscured counterparts of the x - ray detected agns . detailed observations of individual objects are now required to elucidate the properties of this x - ray undetected agn population . alexander , d. m. , et al . 2001 , , 122 , 2156 alexander , d. m. , et al . 2002 , , 568 , l85 alexander , d. m. , et al . 2003 , , 126 , 539 alonso - herrero , a. , et al . 2006 , , 640 , 167 bauer , f. e. , et al . 2004 , aj , 128 , 2048 brand , k. , et al . 2006 , , 644 , 143 brandt , w. n. , et al . 2001 , apj , 558 , l5 brandt , w. n. , & hasinger , g. 2005 , ara&a , 43 , 827 colbert , e. j. m. , et al . 2004 , , 602 , 231 donley , j. l. , rieke , g. h. , rigby , j. r. , & prez - gonzlez , p. g. 2005 , , 634 , 169 efstathiou , a. , & rowan - robinson , m. 1995 , mnras , 273 , 649 elvis , m. , et al . 1994 , , 95 , 1 gebhardt , k. , et al . 2000 , , 539 , l13 george , i. m. , et al . 2000 , , 531 , 52 granato , g. l. , & danese , l. 1994 , , 268 , 235 jarvis , m. j. , & rawlings , s. 2004 , new astronomy review , 48 , 1173 kormendy , j. , & richstone , d. 1995 , , 33 , 581 lacy , m. , et al . 2004 , , 154 , 166 magorrian , j. , et al . 1998 , aj , 115 , 2285 mainieri , v. , et al . 2005 , , 437 , 805 page , m. j. , et al . 2006 , mnras , 369 , 156 piccinotti , g. , et al . 1982 , , 253 , 485 ptak , a. , serlemitsos , p. , yaqoob , t. , & mushotzky , r. 1999 , , 120 , 179 rees , m. j. 1984 , , 22 , 471 stern , d. , et al . 2005 , apj , 631 , 163 tozzi , p. , et al . 2006 , a&a , 451 , 457 treister , e. , & urry , c. m. 2005 , apj , 630 , 115 treister , e. , et al . 2006 , , 640 , 603 wirth , g. d. , et al . 2004 , , 127 , 3121 worsley , m. a. , et al . 2005 , mnras , 357 , 1281 worsley , m. a. , et al . 2006 , , 368 , 1735	deep x - ray surveys provide the most efficient identification of active galactic nuclei ( agn ) activity . however , recent evidence has indicated that the current @xmath0 kev surveys do not detect the most heavily obscured agns . here we explore whether the x - ray undetected agn population can be identified within the ultra - deep _ spitzer _ survey of the goods - n field using x - ray stacking techniques . we find evidence for agn activity in the _ spitzer _ dataset and the strongest and hardest x - ray signal is produced by galaxies with starburst - like infrared spectral slopes and median properties of @xmath1 0.8 and @xmath2 @xmath3 . the stacked x - ray properties suggest that obscured agn activity is present in these sources , with a median x - ray spectral slope of @xmath4 and @xmath5 erg s@xmath6 . these overall properties are consistent with the obscured agn population expected to produce the unresolved x - ray background .
under soc , it is more transparent to re - write the hamiltonian in the helicity basis : @xmath121 where the annihilation operators of the helicity branches @xmath122 , with corresponding dispersion @xmath123 . the interaction between the impurity atom and atoms in the helicity branches is thus momentum dependent . we consider a system with a fixed total particle density , as is the case in experiments . when the total particle density is fixed , changes in the soc parameters @xmath9 or @xmath10 would affect the single particle dispersion , which would change the fermi energy @xmath64 as well . since it is more convenient to vary the intensity of the raman fields experimentally , we focus on the cases where @xmath10 is fixed , while @xmath9 varies . for simplicity , we only consider cases with @xmath67 , for which there is no single - particle ground state degeneracy in the lower helicity branch . there are then two fermi surfaces , one in each helicity branch , for @xmath124 ; and only one fermi surface in the lower branch for @xmath125 . for given parameters , the fermi energy @xmath64 can be determined from the total number density @xmath126 as @xmath127 . the contour of the fermi surface in momentum space can be calculated from the expression @xmath128 , which in turn determines the range of integrations in the closed equations in the main text . a typical evolution of the fermi surface with varying parameters is given in fig . [ fig : e0alp05 ] . under the bcs mean - field treatment in the text , the effective hamiltonian can be written in a quadratic form : @xmath130 where @xmath131 is a 3@xmath1323 matrix : @xmath133 where we have assumed the order parameter @xmath104 to be real . applying the unitary transfer on the spin basis : @xmath134 @xmath135 can be diagonalized : @xmath136 where @xmath137 is the transformation matrix containing three eigen - vectors of @xmath138 , and @xmath139 are the annihilation operators for quasi - particles with energies @xmath140 . then for the zero - temperature ground state we can get the expression of the thermodynamic potential @xmath135 as eq.(7 ) in the text , where the expectation value @xmath141 is taken with respect to the vacuum state of the quasi - particles	we investigate the pairing physics in a three - component fermi - fermi mixture , where a few fermionic impurities are immersed in a non - interacting two - component fermi gas with synthetic spin - orbit coupling ( soc ) , and interact attractively with one spin species in the fermi gas . due to the interplay of soc and spin - selective interaction , the molecular state intrinsically acquires a non - zero center - of - mass momentum , which results in a new type of fulde - ferrell ( ff ) pairing in spin - orbit coupled fermi systems . the existence of the fermi sea can also lead to the competition between ff - like molecular states with different center - of - mass momenta , which corresponds to a first - order transition between ff phases in the thermodynamic limit . as the interaction strength is tuned , a polaron - molecule transition occurs in the highly imbalanced system , where the boundary varies non - monotonically with soc parameters and gives rise to the reentrance of polaron states . the rich physics in this system can be probed using existing experimental techniques . _ introduction_. the realization of synthetic spin - orbit coupling ( soc ) in ultracold atomic gases has triggered a great amount of experimental interests @xcite . in ultracold fermi gases , various forms of soc in different dimensions can give rise to a wealth of exotic superfluid phases @xcite . in particular , recent studies have shown that by deforming the fermi surface in the presence of soc , pairing states with non - zero center - of - mass ( com ) momentum , the so - called fulde - ferrell ( ff ) states , can be stabilized over a large parameter region @xcite . in all these cases , soc proves to be a powerful tool of quantum control , which , when combined with the outstanding tunability of ultracold atomic gases , can often lead to novel pairing superfluidity . in this work , we consider another promising system for exotic pairing states , where a few fermionic impurities are immersed in a non - interacting spin-@xmath0 fermi gas with the synthetic soc that has been realized in cold atoms experiments @xcite . in this three - component mixture , the impurity fermions are tuned close to a wide feshbach resonance with one particular spin species in the two - component fermi gas ( see fig . [ fig : schematic ] ) . we demonstrate that such a system exhibits interesting pairing physics that can be readily probed using the existing experimental techniques . ( see text ) . the pairing states naturally acquire a non - zero com momentum in the system.,width=283 ] a fundamental feature of the system is that the two - body bound state naturally acquires a non - zero com momentum , corresponding to an ff pairing phase in the thermodynamic limit . this can be attributed to the interplay between soc and spin - selective interaction , a mechanism different from that of the previously studied ff phases in spin - orbit coupled systems @xcite . moreover , the existence of the spin - orbit coupled fermi sea can lead to the competition between two ff - like molecules with different com momenta . consequently , we find a first - order transition between two ff pairing phases on the mean - field phase diagram in the thermodynamic limit . for a highly imbalanced system with a single impurity , the ground state can undergo polaron - molecule transitions , corresponding to phase transitions from the normal to the superfluid phase in many - body systems @xcite . interestingly , it is found that the critical interaction strength for such transitions varies non - monotonically with soc parameters , giving rise to the reentrance of polaron states for certain interaction strengths . these properties can be directly probed in fermi - fermi mixtures of @xmath1k-@xmath1k-@xmath1k or @xmath2li-@xmath1k-@xmath1k using existing experimental techniques @xcite . _ model_. the hamiltonian of our system is written as : @xmath3 where @xmath4 ( @xmath5 ) is the annihilation operator for the spin components in the fermi gas , and @xmath6 is the annihilation operator for the impurity atoms . the kinetic energy of the atoms are given as @xmath7 , @xmath8 . the soc parameters @xmath9 and @xmath10 are respectively proportional to the raman coupling strength and the momentum transfer in the raman process generating the soc @xcite . for simplicity , we have assumed a vanishing two - photon detuning . the bare interaction rate @xmath11 between the impurity and the spin - up atom can be renormalized as @xmath12 , where @xmath13 is the quantization volume , @xmath14 is the mass ratio , @xmath15 and @xmath16 is the @xmath17-wave scattering length . for a natural description of the system , we transform the annihilation operators into the helicity basis : @xmath18 , with @xmath19 , and @xmath20^{1/2}/\sqrt{2 } [ h^2+\alpha^2k_x^2]^{1/4}$ ] . the single - particle dispersion of the helicity branches @xmath21 ( see fig . [ fig : schematic ] ) . the momentum dependence of @xmath22 also gives rise to a momentum - dependent effective interaction between impurity atoms and atoms in the helicity branches @xcite . _ two - body states_. we first consider two - body bound states , which already provides key information of pairing states in the system . we adopt the ansatz wave function : @xmath23 where @xmath24 are variational coefficients , @xmath25 is the com momentum of the dimer , and @xmath26 is the vacuum sate . the self - consistent equation for the dimer energy @xmath27 can be obtained by minimizing the energy functional @xmath28 : @xmath29 the ground state lies with the @xmath25-sector that gives the lowest energy solution of eq . ( [ eqn : twobodyeqn ] ) . we find numerically that this always occurs for @xmath25 along the @xmath30-direction . for the two - body calculations , we take the unit of energy @xmath31 , and the unit of momentum @xmath32 . , with @xmath33 ( solid ) , @xmath34 ( dashed ) , @xmath35 ( dash - dotted ) , and @xmath36 ( dotted ) . ( b ) center - of - mass momentum of bound states in ( a ) . ( c ) two - body bound state threshold for varying @xmath9 and a fixed @xmath37 , with the mass ratio @xmath38 ( solid ) , @xmath39 ( dashed ) , and @xmath40 ( dash - dotted ) . ( d ) typical probability distribution of the two - body wave function in the lower helicity branch , @xmath41 , in the @xmath42 plane , with @xmath37 , @xmath33 , @xmath43 , and @xmath38 . the threshold energy @xmath44 and the unit of energy @xmath45 are defined in the text.,width=340 ] in fig . [ fig : twobody](a)(b ) , we show the two - body ground state energy @xmath27 and the com momentum @xmath46 as functions of the interaction strength with @xmath38 . the bound state emerges when @xmath27 is below the threshold energy @xmath44 for two free particles @xcite . the critical interaction strength for its emergence , @xmath47 , is plotted as functions of @xmath9 in fig . [ fig : twobody](c ) . apparently , the presence of synthetic soc pushes the critical @xmath48 toward the bec limit , thus suppressing the formation of two - body bound state . we have also checked that three - body bound states are always metastable in this case , in contrast to the case of a rashba soc @xcite . a general feature of the bound state in our system is that it acquires a finite com momentum ( fig . [ fig : twobody](b ) ) . accordingly , the momentum distribution of the bound state exhibits a peak at finite momentum as shown in fig . [ fig : twobody](d ) . different from the finite - momentum pairing in previous studies under soc and single - particle dispersion asymmetry @xcite , here there is no asymmetry in the single - particle dispersion . instead , the finite - momentum pairing in our system is a combined effect of the spin - selective interaction and the momentum - dependent spin mixture under soc , as schematically shown in fig . [ fig : schematic ] . in the bec limit , the com momentum approaches @xmath49 ( see fig . [ fig : twobody](b ) ) , which is effectively the momentum shift of the spin - up atom , as the system is then dominated by the binding of the impurity atom and the spin - up atom . an important implication of these two - body results is that a new type of ff - like pairing state should exist on the many - body level . _ effects of fermi - sea_. en route to characterizing the many - body pairing in our system , we now consider a single impurity atom interacting with a fermi - sea of helicity atoms , and study the effects of fermi - sea on the pairing physics . in this highly imbalanced system , a polaron - molecule transition is expected to occur , which could be observed experimentally using radio - frequency ( rf ) spectroscopy @xcite . in the thermodynamic limit , the transition should correspond to a superfluid to normal state phase transition . as functions of the center - of - mass momentum @xmath46 for @xmath50 , with @xmath51 ( solid ) , @xmath52 ( dashed ) , and @xmath53 ( dash - dotted ) . ( b ) normalized molecular energies as functions of @xmath46 for @xmath52 , with @xmath54 ( solid ) , @xmath55 ( dashed ) , and @xmath56 ( dash - dotted ) . the normalization in ( b ) is with respect to the molecular energies @xmath57 at @xmath58 . for both figures , @xmath59 , and the mass ratio @xmath40 corresponds to the case of a li - k - k mixture . , width=340 ] _ ( 1 ) molecular state._ the ansatz wave function of the molecular state can be written as : @xmath60 where @xmath61 is the variational coefficient , @xmath62 denotes a spin - orbit coupled fermi sea with @xmath63 atoms and a fermi energy of @xmath64 @xcite . the summation runs over the energy space beyond the fermi surface . minimizing the energy functional , we can get the self - consistent equation for the molecule energy @xmath65 , which is the same as eq . ( [ eqn : twobodyeqn ] ) except for the restrained summation as in eq . ( [ mol_wf ] ) . we find numerically that the ground state always has a com @xmath25 along the @xmath30-direction . for the numerical calculations , we adopt the unit of energy as the fermi energy @xmath66 of a two - component , non - interacting fermi gas in the absence of soc and with the same total number density . for simplicity , we only consider cases with @xmath67 , for which there is no single - particle ground state degeneracy in the lower helicity branch . here the fermi wave vector @xmath68 is defined through @xmath69 . the presence of the fermi sea has significant impacts on the molecular state . due to the pauli blocking of atom - scattering inside the fermi sea , we find that the molecular state becomes more difficult to form compared to the pure two - body case , and typically requires a stronger critical interaction strength . more importantly , the presence of fermi surfaces offer possibilities of pairing at different com momenta , with dominating contributions from either the lower or the upper helicity branch , which can respectively give rise to negative or positive pairing momentum along @xmath70 ( see fig . [ fig : schematic ] ) . indeed , we find that for appropriate parameters , especially for a light impurity atom with a large @xmath71 , there exists a competition between two ff - like molecular states with different com momenta . as shown in fig . [ fig : molemvsq ] , first - order transitions between different ff - like molecular states can occur with varying @xmath9 or @xmath72 , while the number of atoms in the fermi sea is fixed by @xmath73 @xcite . this strongly suggests a first - order transition between different ff pairing phases in the thermodynamic limit . , and different @xmath72 , where @xmath74 ( solid ) , @xmath75 ( dashed ) , and @xmath76 ( dash - dotted ) . the thin red curves represent polaron energy @xmath77 , while the thick blue curves represent the molecular energy relative to the fermi surface @xmath78 . the vertical dotted line marks the boundary between parameter regions with one fermi surface ( 1fs ) and with two fermi surfaces ( 2fs ) . ( b ) boundary of polaron - molecule transition , with @xmath79 ( solid ) , @xmath59 ( dashed ) , @xmath80 ( dash - dotted ) . in all cases , the mass ratio @xmath38.,width=340 ] _ ( 2 ) polaron - molecule transition_. the stability of molecular state is challenged by the polaron state , which is characterized by particle - hole excitations above the fermi - sea . up to the lowest order excitation , the polaron wave function is written as : @xmath81 where @xmath82 ( @xmath83 ) denotes the two helicity branches , @xmath84 and @xmath85 are the variational coefficients . minimizing the energy functional , we have the equation for the polaron energy @xmath86 : @xmath87 here we have considered a polaron state with zero com momentum @xcite . in fig . [ fig : polmol](a ) , we compare the polaron energy @xmath86 with @xmath88 , the molecular energy relative to a spin - orbit coupled fermi sea of @xmath89 atoms , for various interaction strengths as @xmath9 varies . interestingly , the trajectories of these two energies behave differently when the number of fermi surfaces changes from two to one as @xmath9 increases . consequently , for appropriate parameters , two separate transition points can be crossed by tuning @xmath9 ( see dash - dotted curves in fig . [ fig : polmol](a ) ) . in fig . [ fig : polmol](b ) , we plot the polaron - molecule transition boundaries as functions of @xmath9 . one can see clearly the non - monotonic behavior of the critical @xmath48 , with a minimum kink appearing when the fermi surface just touches the upper helicity branch . thus for certain fixed interaction strengths , the system can go through two transition boundaries by changing the soc parameter @xmath9 , leading to an exotic reentrance of the polaron state . this reentrance phenomenon has never been found in previous studies of polaron - molecule transition in fermi gases . _ many - body phase diagram_. the rich physics discussed above should also be manifest in the pairing phases and phase transitions on the many - body level . to demonstrate this , we study the pairing superfluidity of the system using the standard bcs - type mean - field theory , where the thermodynamic potential at zero temperature @xmath90 becomes @xcite : @xmath91 here , @xmath92 , @xmath93 is the quasi - particle energies , the order parameter @xmath94 , @xmath95 ( @xmath96 ) is the chemical potential of the corresponding atoms , @xmath97 ( @xmath98 ) is the corresponding particle number , and @xmath99 is the heaviside step function . given the chemical potentials , the ground state can be obtained by minimizing @xmath100 with respect to @xmath101 and @xmath102 . numerically , we find that the ground state @xmath101 for the ff pairing always lies along the @xmath30-direction . ( solid , left @xmath103-axis ) and pairing gap @xmath104 ( dashed , right @xmath103-axis ) as functions of @xmath105 . here we choose @xmath106 . ( b ) phase diagram in the ( @xmath107 ) plane . the red ( blue ) curve shows the n - ff1 ( ff1-ff2 ) transition boundary , where both boundaries are of first order . @xmath9 and @xmath95 are the same as in ( a)(b ) . in all cases , the mass ratio @xmath40.,width=340 ] in fig . [ fig : phase](a ) , we show how the ground state @xmath46 and @xmath104 evolve with the interaction parameter @xmath105 , with other parameters ( @xmath108 ) fixed . one sees clearly that by increasing @xmath105 , the system first goes through a first - order phase transition from the normal phase ( n ) to an ff pairing phase ( ff1 ) , and then another first - order transition to a new ff phase ( ff2 ) with different @xmath109 and @xmath104 . the emergence of multiple ff phases and the first - order boundaries between them are qualitatively consistent with our previous analysis on molecular states , though the actual values of @xmath46 and @xmath110 at the phase boundary are now affected by the existence of the impurity fermi sea . in fig . [ fig : phase](b ) , we show the phase diagram when the chemical potential @xmath96 is further tuned . again , over a wide region of @xmath111 , the system can be tuned across all three phases ( n - ff1-ff2 ) by adjusting the interaction strength . hence , these exotic phases and phase transitions can be probed experimentally by sweeping the magnetic field across a feshbach resonance of the spin - up and the impurity atoms . _ summary_. in summary , we have systematically studied the pairing physics in a three - component fermi - fermi mixture , where two constituent fermions are subject to synthetic spin - orbit coupling . we show that the ground state pairing of this system intrinsically features a finite com momentum , which originates from a distinctively new ff pairing mechanism in spin - orbit coupled fermi systems . the intriguing properties of the two - body bound state , the polaron - molecule transition and the ff paring superfluidity in this system can be explored in current cold atom experiments on @xmath1k-@xmath1k-@xmath1k or @xmath2li-@xmath1k-@xmath1k mixtures . for example , in a @xmath1k-@xmath1k-@xmath1k mixture , a raman process with a coupling strength of @xmath112 and a fermi energy of @xmath113 should lead to @xmath114 and @xmath115 . here the recoil energy @xmath116 ( @xmath117 ) is typically on the order of @xmath118khz for @xmath1k @xcite . under these parameters , the polaron - molecule transition occurs at @xmath119 . by further varying the parameters @xmath120 , or by choosing different atomic species and adjusting their numbers , many of the interesting phenomena discussed in this work should emerge given low enough temperatures . _ we thank jing zhou for her early contribution to this project . this work is supported by nfrp ( 2011cb921200 , 2011cba00200 ) , nnsf ( 60921091 ) , nsfc ( 11104158,11374177,11105134,11374283 ) , the fundamental research funds for the central universities ( wk2470000006 ) , and the programs of chinese academy of sciences . 99 y .- j . lin , k. jimnez - garca , and i. b. spielman , nature ( london ) * 471 * , 83 ( 2011 ) . j .- y . zhang , s .- c . ji , z. chen , l. zhang , z .- d . du , b. yan , g .- s . pan , b. zhao , y .- j . deng , h. zhai , s. chen , and j .- w . pan , phys . rev . lett . * 109 * , 115301 ( 2012 ) . j .- y . zhang , s .- c . ji , l. zhang , z .- d . du , w. zheng , y .- j . deng , h. zhai , s. chen , and j .- w . pan , arxiv:1305.7054 . p. wang , z .- q . yu , z. fu , j. miao , l. huang , s. chai , h. zhai , and j. zhang , phys . rev . lett . * 109 * , 095301 ( 2012 ) . l. w. cheuk , a. t. sommer , z. hadzibabic , t. yefsah , w. s. bakr , and m. w. zwierlein , phys . rev . lett . * 109 * , 095302 ( 2012 ) . z. fu , l. huang , z. meng , p. wang , l. zhang , s. zhang , h. zhai , p. zhang , and j. zhang , nat . phys . 2824 ( 2013 ) . h. zhai , int . j. mod . phys . b * 26 * , 1230001 ( 2012 ) . c. zhang , s. tewari , r. m. lutchyn , and s. das sarma , phys . rev . lett . * 101 * , 160401 ( 2008 ) . m. sato , y. takahashi , and s. fujimoto , phys . rev . lett . * 103 * , 020401 ( 2009 ) . j. p. vyasanakere , s. zhang , and v. b. shenoy , phys . rev . b * 84 * , 014512 ( 2011 ) . m. gong , s. tewari , and c. zhang , phys . rev . lett . * 107 * , 195303 ( 2011 ) . h. hu , l. jiang , x .- j . liu , and h. pu , phys . rev . lett . * 107 * , 195304 ( 2011 ) . z .- q . yu and h. zhai , phys . rev . lett . * 107 * , 195305 ( 2011 ) . m. iskin and a. l. subasi , phys . rev . lett . * 107 * , 050402 ( 2011 ) . w. yi and g .- c . guo , phys . rev . a * 84 * , 031608(r ) ( 2011 ) . l. dellanna , g. mazzarella , and l. salasnich , phys . rev . a * 84 * , 033633 ( 2011 ) . m. gong , g. chen , s. jia , and c. zhang , phys . rev . lett . * 109 * , 105302 ( 2012 ) . l. han and c. a. r. s de melo , phys . rev . a * 85 * 011606(r ) ( 2012 ) . r. liao , y. yi - xiang , and w .- m . liu , phys . rev . lett . 108 , 080406 ( 2012 ) . l. he and x .- g . huang , phys . rev . lett . * 108 * , 145302 ( 2012 ) . j. zhou , w. zhang , w. yi , phys . rev . a * 84 * , 063603 ( 2011 ) . x. yang and s. wan , phys . rev . a * 85 * , 023633 ( 2012 ) . l. dong , l. jiang , and h. pu , new j. phys . * 15 * , 075014 ( 2013 ) . x .- f . zhou , g .- c . guo , w. zhang , and w. yi , phys . rev . a * 87 * , 063606 ( 2013 ) . x .- j . liu , k. t. law , t. k. ng , phys . rev . lett . * 112 * , 086401 ( 2014 ) . p. fulde and r. a. ferrell , phys . rev . * 135 * , a550 ( 1964 ) ; a. i. larkin and y. n. ovchinnikov , sov . phys . jetp * 20 * , 762 ( 1965 ) . k. michaeli , a. c. potter , and p. a. lee , phys . rev . lett . * 108 * , 117003 ( 2012 ) . v. b. shenoy , phys . rev . a * 88 * , 033609 ( 2013 ) . l. dong , l. jiang , h. hu , and h. pu , phys . rev . a * 87 * , 043616 ( 2013 ) . z. zheng , m. gong , x. zou , c. zhang , and g .- c . guo , phys . rev . a * 87 * , 031602(r ) ( 2013 ) . f. wu , g .- c . guo , w. zhang , and w. yi , phys . rv . lett . * 110 * , 110401 ( 2013 ) . c. chen , phys . rev . lett . * 111 * , 235302 ( 2013 ) . c. q , z. zhen , m. gong , y. xu , l. mao , x. zhou , g. guo , and c. zhang , nat . comm . * 4 * , 2710 ( 2013 ) . w. zhang and w. yi , nat . comm . * 4 * , 2711 ( 2013 ) . x .- j . liu and h. hu , phys . rev . a * 88 * , 023622(r ) ( 2013 ) . f. chevy , phys . rev . a * 74 * , 063628 ( 2006 ) . r. combescot and s. giraud , phys . rev . lett . * 101 * , 050404 ( 2008 ) . s. zllner , g. m. bruun , c. j. pethick , phys . rev . a * 83 * , 021603(r ) ( 2011 ) m. m. parish , phys . rev . a * 83 * , 051603 ( r ) ( 2011 ) . a. schirotzek , c .- h . wu , a. sommer , and m. w. zwierlein , phys . rev . lett . * 102 * , 230402 ( 2009 ) . m. koschorreck , d .. pertot , e. vogt , b. frhlich , m. feld , and m. khl , nature * 485 * , 619 ( 2012 ) . w. yi and w. zhang , phys . rev . lett . * 109 * , 140402 ( 2012 ) . see supplementary materials for the helicity basis hamiltonian and the details of solving two - body and many - body pairing problems . x. cui and w. yi , arxiv:1403.0649 . j. t. stewart , j. p. gaebler , and d. s. jin , nature * 454 * , 744 ( 2008 ) . it is straightforward to show numerically that the ground state in the polaron sector features a very small nonzero center - of - mass momentum , and a very small decrease in energy .
it is well known that one of the most important indicators of the chaotic behaviour of a dynamical system is the rate of decay to zero of the time autocorrelation of dynamical variables . a vast literature exists on this subject , mainly addressed to systems of some special class ( for example , anosov systems , see @xcite , or systems with few degrees of freedom , such as billiards , see @xcite ) . in this paper we study the decay of correlations in the general frame of the dynamical theory of hamiltonian systems , with an invariant measure . such an approach has already provided interesting results , as it allows to obtain results valid in the thermodynamic limit ( see @xcite ) . here we provide a power series expansion ( with respect to time ) of the time autocorrelation of a smooth function @xmath0 ( theorem [ teor : serie ] , section [ sez : generale ] ) . it turns out that the coefficients of such a series are essentially the variances of @xmath1 , where @xmath2 is the poisson bracket operator relative to the hamiltonian @xmath3 . we then establish a necessary and sufficient condition for its time - decay not to be exponential ( corollary [ cor : analitico ] , section [ sez : serie ] ) . we finally give a numerical application to the case of the fermi - pasta - ulam system ( fpu ) . by truncating the series up to order 12 , for systems up to 364 degrees of freedom , we approximate the poles of the laplace transform of the time autocorrelation of a suitably chosen dynamical variable . some comments are made on the possibility that such approximations give indications for a sub - exponential decay of the autocorrelation . in section [ sez : generale ] the relevant notions on time correlation functions are recalled . in the same section we give a theorem which establishes a kind of continuity ( with respect to a suitable distance ) of time correlations in the @xmath4 space of the dynamical variables , and also theorem [ teor : serie ] for the series expansion . section [ sez : serie ] is devoted to a study of some general properties of the series expansion . here an interesting link with the stieltjes moment problem is pointed out ( see theorem [ teor : stieltjes ] ) , and a necessary and sufficient criterion for the time - decay of the autocorrelation to be exponential is given as corollary [ cor : analitico ] . a closer examination of this problem is given in appendix [ app : hausdorff ] . then , the results of some numerical computations on the time autocorrelations of a significant variable in a fpu chain are reported in section [ sez : numerica ] . finally , in section [ sez : conclusione ] some comments on the usefulness of the present method are also given . two more appendices complete the paper . we recall here some standard concepts within the measure theoretic approach to dynamical systems . we consider a hamiltonian system on a phase space @xmath5 endowed with a probability measure @xmath6 invariant with respect to the time flow @xmath7 induced by the hamiltonian @xmath3 . we also consider the time evolution operator @xmath8 acting on the space of the square integrable functions from @xmath5 to @xmath9 , i.e. , @xmath10 . the operator @xmath8 maps @xmath0 to @xmath11 . it is known after koopman ( see @xcite ) that @xmath12 defines a one - parameter group of unitary operators , namely operators preserving the norm @xmath13 in @xmath10 . we are interested in the time autocorrelation of a dynamical variable @xmath14 , which is defined as @xmath15 where @xmath16 is a shorthand for @xmath17 , and @xmath18 denotes mean value with respect to the probability measure @xmath6 . it will also be useful to consider the time correlation between two dynamical variables @xmath0 and @xmath19 , defined as @xmath20 it is well known from probability theory that the concepts of variance and correlation acquire a geometrical meaning if the covariance between two random variables @xmath0 and @xmath19 is used as a scalar product . the covariance is defined as @xmath21 this choice leads to take as norm actually coincides with the usual @xmath4 norm . in this paper we shall use both norms , keeping the usual symbol @xmath22 for the @xmath4 norm . ] of a dynamical variable @xmath0 its standard deviation @xmath23 , defined by @xmath24 @xmath25 being the variance of @xmath0 . the notion that some relation exists between two random variables @xmath0 and @xmath19 is made quantitative through the correlation coefficient @xmath26 defined as @xmath27 using the schwarz inequality it is easily seen in our case that one has @xmath28 , so that @xmath29 . two variables are orthogonal in this metric if they are uncorrelated , i.e. , if @xmath30 , and collinear if @xmath31 . we also emphasise that , in view of the above definitions , the time autocorrelation of a function @xmath0 is nothing but the covariance between @xmath32 and @xmath0 . therefore one immediately has @xmath33 thus , the variance of a dynamical variable is the natural scale of its time autocorrelation . 1ex we come now to state two useful theorems . the first one points out an interesting property of the time correlation . the mere fact that two dynamical variables are strongly correlated ( i.e. , sufficiently collinear in the scalar product given by the covariance ) entails a similar behaviour of their time autocorrelations . equivalently , we can say that in the neighbourhood of each dynamical variable @xmath0 , i.e. , in the set of all dynamical variables strongly correlated with @xmath0 , the behaviour of the time autocorrelations is determined by that of @xmath34 . [ teor : correlate ] let @xmath6 be a probability measure on the phase space @xmath5 , invariant for the flow generated by @xmath3 , and let @xmath0 and @xmath19 be dynamical variables belonging to @xmath10 such that one has @xmath35 for some @xmath36 . then there exists a multiple @xmath37 of @xmath0 , with @xmath38 , such that @xmath39 and @xmath40 * proof . * both inequalities come from the remark that @xmath41 which is due to the identities @xmath42 and @xmath43 . in fact , ( [ eq : maggiorazione_correlazione ] ) hence follows by noting that @xmath44 while , in a similar way , ( [ eq : maggiorazione_autocorrelazione ] ) comes from the following relations @xmath45 here , in the second line , use is made of the previous inequality and of the identity @xmath46 , due to the invariance of the measure . 1ex a central role in this paper will be played by the formal power series expansion of @xmath34 with respect to time , which is given in the following theorem . use will be made of the definition of the @xmath47-th order lie derivative of @xmath0 , for @xmath48 , namely @xmath49 $ ] , where @xmath50 and @xmath51 $ ] denotes the poisson brackets . [ teor : serie ] let @xmath6 be a probability measure on the phase space @xmath5 , invariant for the flow generated by @xmath3 , and let @xmath52 . then , for any dynamical variable @xmath14 such that @xmath53 for all @xmath54 , one has @xmath55 * proof . * we use the simple chain of identities @xmath56 which hold true for any invariant measure , and the remark that @xmath57 where @xmath58 $ ] . one can check equations ( [ eq : correlazione_inversa ] ) by writing down the square at the l.h.s . , while ( [ eq : evoluzione_differenza ] ) comes from the variation of constants formula applied to @xmath59 . we go on by writing the l.h.s . of ( [ eq : correlazione_inversa ] ) as @xmath60 where , in the second line , the invariance of the measure with respect to the time evolution is used , so that the equality proceeds from a simple change of coordinates , and in the last line the same argument is used , as well as relation ( [ eq : correlazione_inversa ] ) applied to the variable @xmath61 . the order of the integrals can be exchanged if @xmath62 is finite , because the integral ( [ passo_1 ] ) is absolutely convergent in this case , in virtue of schwarz inequality . in ( [ passo_2 ] ) we repeat the same calculation for @xmath63 , using @xmath64 $ ] . this gives us the relation @xmath65 where @xmath66 $ ] . the same procedure can be iterated up to @xmath67 by recursively defining @xmath49 $ ] , starting with @xmath50 , and using the hypothesis that @xmath68 . this completes the proof . 1ex * remark 1 . * theorem [ teor : serie ] suggests that , at least in a formal way , one can express the time autocorrelation of a dynamical variable @xmath0 as a power series with respect to time @xmath69 to simplify the notations , we define @xmath70 we point out that all these coefficients are positive . we could have got to formula ( [ eq : serie ] ) also by expressing @xmath16 as a time power series through its lie derivatives and by integrating @xmath71 by parts , taking into account the observation that the mean value of any function which can be written as @xmath72 $ ] vanishes can replace @xmath73 in all the previous formulae . we will keep the given notation , because this way it is more straightforward to understand how to compute the coefficients of the series . ] , since @xmath74\bigr\rangle=\frac{{\mathrm{d}}}{{\mathrm{d}}t } \langle g\rangle$ ] . 1ex * remark 2 . * we emphasise that the term in the second line in equation ( [ eq : serie_troncata ] ) , i.e. , the remainder , turns out to be positive or negative according to @xmath75 being odd or even , respectively . therefore , the truncations of series ( [ eq : serie ] ) provide upper bounds for the time autocorrelation of @xmath0 if truncated at an odd @xmath75 , and lower bounds if truncated at an even @xmath75 . such bounds provide some information on the behaviour of the time autocorrelation of @xmath0 for finite times . the simplest example is that of the first order truncation , which proved to be very helpful for the study of relaxation times in a hamiltonian setting ( see @xcite ) . 1ex as a final remark , we point out that some _ a priori _ restrictions on the norms of the successive derivatives come from the schwarz inequality ( [ eq : correlazione_e_varianza ] ) . indeed , one has the following proposition , whose proof is deferred to appendix [ app : a_priori ] . [ prop : a_priori ] the coefficients @xmath76 defined by ( [ eq : definizioni_a_e_c ] ) are such that the polynomials @xmath77 , @xmath78 defined as @xmath79 with @xmath80 are positive semi - definite , for any @xmath81 . the statement of proposition [ prop : a_priori ] entails some restrictions on the coefficients @xmath76 , since @xmath77 and @xmath78 are two - indexed sequences of positive semi - definite polynomials . unfortunately , such relations can be expressed only in a quite complicated way , through the jacobi - borchardt theorem ( see , for example , @xcite ) , which gives some relations among the roots of the polynomials , and newton s identities on the relations between the coefficients of a polynomial and its roots . in many cases one is interested in determining whether the time correlation tends to zero as @xmath82 and , if this is the case , which is the law controlling the decay . indeed , it is commonly stated that if the correlations decay exponentially fast ( for a suitable wide class of functions ) then the system is chaotic . the decay is linked to the analyticity property of the fourier transform of the correlations ( see for example @xcite ) : in particular it is well known that the fourier transform is analytic in a strip of width @xmath83 if and only if the correlation decays at least as @xmath84 . we recall that the correlations can always be expressed as the fourier transform of an even positive measure , the so called `` spectral measure '' , namely as @xmath85 where @xmath86 is a positive borel measure on @xmath87 ( see , for example , @xcite ) . on the other hand , it is not so easy to get information about the spectrum : we devise to get it using the laplace transform @xmath88 of the time autocorrelation , namely , @xmath89 it is well known that the laplace transform of a function which decays faster than @xmath84 is analytic in the half - plane @xmath90 , so that , again , the control on the decay of correlations in linked to the analyticity properties of @xmath88 , as in the case of the fourier transform . the great advantage of the latter approach lays in the fact that the laplace transform of the correlation is essentially the stieltjes transform of the spectral measure . this is stated as theorem [ teor : stieltjes ] below . the point is that , as first shown by stieltjes ( see @xcite ) , if one knows the asymptotic expansion in powers of @xmath91 of a functions @xmath88 which is the stieltjes transform of a positive measure , then one can recover @xmath88 itself through a convergent scheme of continued fractions expansions , defined in terms of the coefficients of the asymptotic expansion . in other words , one can construct a rational approximation @xmath92 which converges uniformly in @xmath93 , as @xmath94 , to @xmath88 . for the case of the time autocorrelation @xmath34 , the asymptotic expansion is obtained just by integrating the expansion term by term . in fact , the laplace transform can be formally written as @xmath95 this expansion would give a convergent expansion , provided one can exchange the sum of the series with the integration over time . this is in general not permissible , in view of the fact that the expansion in power series of time converges ( in general ) only in a circle of finite radius and not up to infinity . however , it is easy to show that series ( [ eq : trasformata_serie ] ) nevertheless provides the asymptotic expansion of @xmath88 we are looking for . one has to use formula and check that the remainder grows at most as a power of time . we are now able to state the main result of this section , namely , the following theorem , together with a relevant corollary on the decay of the correlations . [ teor : stieltjes ] let @xmath34 be analytic in @xmath96 about the origin and continuous for any @xmath97 . then the following statements hold : 1 . in the half - plane @xmath98 the laplace transform @xmath88 of @xmath34 is analytic and one has @xmath99 where @xmath100 is the positive borel measure such that holds ; 2 . [ prop : stieltjes_2 ] the continued fraction which approximates the asymptotic series ( [ eq : trasformata_serie ] ) converges to @xmath88 for @xmath98 and the borel positive measure @xmath101 defined on @xmath102 by @xmath103 solves the stieltjes moment problem for the coefficients @xmath76 defined by ( [ eq : definizioni_a_e_c ] ) . moreover , @xmath104 solves the symmetric hamburger moment problem with moments @xmath105 , @xmath106 . [ cor : analitico ] @xmath34 decays exponentially fast as @xmath96 goes to infinity if and only if the coefficients @xmath76 are such that @xmath107 exists and @xmath108 is analytic . * * remark 3.*we will state our results in terms of the regularity properties of @xmath86 or @xmath109 according to convenience . we prefer the study of the properties of @xmath109 , when the convergence properties of stieltjes continued fraction approximation are needed . 1ex * proof of theorem [ teor : stieltjes ] . * * first of all , we observe that ( [ eq : stieltjes_laplace ] ) follows by expressing @xmath34 via ( [ eq : bochner ] ) in the definition ( [ eq : trasformata_funzione ] ) of @xmath88 and exchanging the order of integration . this can be done for any @xmath110 for which @xmath111 , on account of the tonelli - fubini theorem for the stieltjes integral ( see @xcite ) , and it is well known that the so defined @xmath88 is analytic in the same region . we come now to the proof of statement [ prop : stieltjes_2 ] . we notice that the form of series ( [ eq : trasformata_serie ] ) recalls the stieltjes work @xcite on the link between the asymptotic series , their expansion in continued fractions and the stieltjes transform . in dealing with this subject he met with the moment problem that now bears his name . we recall that , given a sequence of positive numbers @xmath76 , the stieltjes moment problem consists in checking whether there exists a measure @xmath109 on @xmath87 such that @xmath112 the solution exists if and only if @xmath113 where the matrices @xmath114 , @xmath115 are defined by @xmath116 \ , \ \tilde{\delta}_n\!{\buildrel { \rm def } \over { = } } \ ! \left[\ ! \begin{array}{cccc } c_1 & c_2 & \cdots & c_{n+1}\\ c_2 & c_3 & \cdots & c_{n+2}\\ \vdots & \vdots & \ddots & \vdots\\ c_{n+1 } & c_{n+2 } & \cdots & c_{2n+1 } \end{array } \!\right]\ , \ ] ] and if the equality in ( [ eq : soluzione_stieltjes ] ) is obtained for @xmath117 , then it is obtained for @xmath114 , with @xmath118 and for @xmath115 , with @xmath119 . furthermore , the solution is unique if there exists @xmath120 such that @xmath121 holds ( for a proof with more recent methods , see @xcite ) . we point out that , in our case , since @xmath122 is analytic about the origin , there should exist @xmath120 such that the previous condition is satisfied , in view of the cauchy - hadamard criterion . coming back to our problem , we collect the results of paper @xcite of interest to us in the following statement : _ if the function @xmath88 admits the representation ( [ eq : stieltjes_laplace ] ) , then the continued fraction which approximates the asymptotic expansion ( [ eq : trasformata_serie ] ) converges to @xmath88 for @xmath93 ; moreover , the stieltjes moment problem for the sequence @xmath76 is soluble . _ the previous statement is proved this way . in equation ( [ eq : stieltjes_laplace ] ) we put @xmath123 , @xmath124 , @xmath125 . then we have the formal chain of equalities @xmath126 where again @xmath109 is a borel positive measure on @xmath102 . stieltjes showed that the continued fraction which approximates the term in brackets at the r.h.s . converges to the integral at the l.h.s , and that the stieltjes moment problem for the sequence @xmath76 has a solution for the measure @xmath109 . then it is straightforward to check that @xmath127 solves the associated symmetric hamburger moment problem , which means that @xmath128 q.e.d . 1ex * remark 4 . * notice that we have incidentally shown that the coefficients @xmath76 are such that conditions ( [ eq : soluzione_stieltjes ] ) are satisfied at any order . this restriction on the values of the coefficients , together with the analogous ones imposed by the application of the previous theorem to the dynamical variables @xmath129 , for all @xmath75 , already contains the restriction given by proposition [ prop : a_priori ] . however , we point out that proposition [ prop : a_priori ] holds up to a given order even if the existence of @xmath129 beyond such an order is not guaranteed . 1ex in view of corollary [ cor : analitico ] , the question of the asymptotic behaviour of the time autocorrelations can be restated as a regularity problem for the measure @xmath109 , which solves stieltjes moment problem for the @xmath76 , or , equivalently , for @xmath86 which solves the associated hamburger moment problem . an answer to this question can be given only if the whole sequence of moments @xmath76 is known . however , it is also of interest to understand which is the qualitative difference between the approximation ( at a given order ) of an analytic measure and that of a measure with at least a singular point . we try here to give an answer to the latter question , deferring to appendix [ app : hausdorff ] the discussion of which is the link between the behaviour of the moments and the regularity of the solution ( see proposition [ prop : ahiezer ] , [ prop : hausdorff ] , [ prop : convergente ] therein ) . we have already pointed out that a rational approximation @xmath130 of the laplace transform @xmath88 can be constructed from ( [ eq : sviluppo_serie ] ) by standard methods ( see chapter 2 of stieltjes memoir @xcite ) , which give @xmath131 such a function , in turn , can be rewritten as the following stieltjes integral @xmath132 in which the measure @xmath133 is piecewise constant and has a jump @xmath134 at the points @xmath135 . therefore , in order that the limiting measure @xmath136 be continuous , it is necessary that , as the approximation order grows , the poles become dense and their residues tend to zero . as a consequence , _ we will take as a qualitative indication of a sub - exponential decay the property that the rational approximations have at least a pole which remains isolated , with a stable residue , as the order grows_. checking whether the spectral measure is analytic , or even whether the criterion of proposition [ prop : ahiezer ] of appendix [ app : hausdorff ] is satisfied for a concrete problem is a major task , because it requires information on an infinite number of coefficients . thus , as a first attempt to implement the ideas of section [ sez : serie ] , we computed numerically some coefficients @xmath76 for a model which is widely studied in the literature and is of great relevance for the foundations of statistical mechanics , namely the fermi - pasta - ulam model ( fpu , see @xcite ) . it describes a one - dimensional chain of @xmath137 particles interacting through nonlinear springs . the hamiltonian of such a system , after a suitable rescaling , can be written as @xmath138\ , \ ] ] where @xmath139 and @xmath140 are canonically conjugated variables , while @xmath86 and @xmath141 are parameters that control the size of the nonlinearity . we choose as invariant probability measure the gibbs one , i.e. , @xmath142 , where @xmath143 is the partition function and @xmath144 the temperature , and for the purpose of speeding up the numerical evaluation of the integrals involving the measure , we consider here a chain with one end point fixed and the other one free , i.e , we impose only the boundary condition @xmath145 , @xmath146 ( see footnote [ nota : indipendenti ] below ) . we note however that a few computations made on the model in which both end - points are fixed have shown no significant difference . concerning the parameters , we consider the often studied case in which @xmath147 . the relevant fact is that as @xmath148 approaches 0 the contribution due to the nonlinear terms becomes statistically negligible . it is well known that if the nonlinear terms are neglected then the hamiltonian is integrable , admitting @xmath149 normal modes of oscillation , and that the motion is quasi periodic so that , for every @xmath150 , the energy @xmath151 of the @xmath47-th mode is a constant of motion ( and thus , as obvious , the time autocorrelation of such a dynamical variable is constant ) . our aim is to investigate what happens when the nonlinearity is introduced as a perturbation , namely , for @xmath148 positive but close to 0 . we are particularly interested in the exchange of energy among the normal modes . to this end , a good choice could be to consider the fraction of energy localized on the low frequency modes , defined , for example , as @xmath152 since its time variation is a convenient indicator of the flow of energy from low to high frequency modes , and vice versa . as a numerical tool we want to compute numerically the coefficients of the series ( [ eq : serie ] ) for this function , since , in view of theorem [ teor : stieltjes ] , they give indications on the asymptotic behaviour of the autocorrelation . as a matter of fact , we notice that the function @xmath153 may not represent the best choice . for , such a function is strongly correlated with the hamiltonian , so that , in virtue of theorem [ teor : correlate ] , its autocorrelation remains close to that of the hamiltonian , thus remaining far from zero . so , we rather took a suitable modification of @xmath153 , namely , the projection of @xmath153 , via the gram - schmidt orthogonalization process , on the space of the dynamical variables uncorrelated with @xmath3 . thus we consider the dynamical variable @xmath154 . our calculation proceeds as follows . we extract a sample in phase space according to the gibbs measure and @xmath155 , for @xmath156 , which are stochastically independent variables for the gibbs measure with respect to the present hamiltonian . here lies the great advantage of studying the chain with one end - point free.[nota : indipendenti ] ] and we estimate the @xmath4-norm @xmath157 of the functions @xmath158 $ ] generated by @xmath159 taking the mean value on our sample . for this purpose , we observe that @xmath160 is a certain function of @xmath161 and @xmath162 , which can be expressed via fa di bruno s formula for the derivatives of any order of a composed function . this formula requires the computation of some combinatorial coefficients , which needs a smart implementation in order to reduce the computation time . we report in appendix [ app : faa ] the scheme we have followed . in order to study the laplace transform of the time autocorrelation of @xmath163 , following the procedure of stieltjes , we approximate it up to order @xmath75 with rational functions , and look for the poles of such an approximating function in the complex plane ( see chapter 2 of the memoir @xcite ) . our aim is to compare the behaviour of the time autocorrelation of @xmath163 with some functions which behave in a somehow known way , by comparing the poles of the corresponding rational approximations of the laplace transforms . the first comparison function is the time autocorrelation of a variable which should lose correlation in a very short time , i.e. , ( the orthogonal projection of ) the kinetic energy @xmath164 of the first half of the chain s particles . the second one will be a given function of time which has an exponential decay ( see below for its actual form ) . inverse of the imaginary part of the four poles of the laplace transform of the time autocorrelation of @xmath165 versus temperature , in logarithmic scale . here the number of particles is @xmath166.,scaledwidth=75.0% ] residues at the four poles , whose imaginary parts are reported in fig . [ fig1 ] versus temperature , in logarithmic scale . here the number of particles is @xmath166.,scaledwidth=75.0% ] let us describe the results in some detail . we know that the rational approximation of the laplace transform at order @xmath167 will have @xmath167 complex conjugated poles , @xmath168 , @xmath169 say . the inverses of the @xmath170 s so found for the time autocorrelation of @xmath171 are plotted in fig . [ fig1 ] versus temperature @xmath148 . the pole that shows up at order 1 produces the upper curve in the figure . it corresponds to oscillation of a quite large period , growing as @xmath172 as @xmath148 decreases to zero . the actual value changes a little when the approximation increases , keeping however the same order of magnitude . the plotted data correspond to the approximation order @xmath173 . we have also noted that the values found do not significantly change with the number @xmath149 of particles , for @xmath149 up to 364 . the frequencies represented by the lower curves show up at successive orders @xmath174 , respectively , and correspond to oscillations of a short period , which , at variance with the first period , are of the order of magnitude of the typical periods of the normal modes . they appear to remain almost constant with @xmath148 . a further information is provided by the residues of the poles , which represent the amplitudes of the oscillations . the quantities for the four poles of fig . [ fig1 ] are reported in fig . [ fig2 ] , after a normalization defined by setting the sum of the residues equal to one . the remarkable fact is that the amplitude corresponding to the longest period is the largest one . even for @xmath175 the amplitudes of the shorter periods do not exceed one percent of the total it is now of interest to look at the corresponding graphs ( fig . [ fig3]-[fig4 ] ) for the time autocorrelation of the kinetic energy of half of the chain , @xmath176 . as expected , the time - scales involved are much shorter , as is seen by inspection of fig . in any case , an indication of a bigger chaoticity is given by the fact that all the residues are of the same order of magnitude and are closer to each other , as can be seen in fig . this means that in the spectrum of the time autocorrelation of @xmath164 at least three frequencies are excited , at variance with the spectrum of @xmath171 , for which just a frequency is actually excited . then , we compare the behaviour of @xmath177 with that of @xmath178 , where @xmath179 , in order to look for possible differences with respect to an exponential decay . here the constant @xmath180 is so chosen that the approximation of the laplace transform of such a function at order @xmath181 has a pole at the same place as the laplace transform of @xmath182 , for a fixed temperature @xmath148 ( as we have seen , @xmath183 ) . in fig . [ fig5 ] the poles and the corresponding residues of the laplace transforms of @xmath184 and @xmath182 are plotted at the successive orders of approximation @xmath185 . residues of the poles of the laplace transforms of @xmath186 and of @xmath187 versus their imaginary parts , in logarithmic scale . the circles refer to @xmath188 and the triangles to @xmath187 , while the dashed lines indicate that the order of approximation is 3 , and the continuous lines indicate order 4 . here the number of particles is @xmath166 and @xmath189,scaledwidth=80.0% ] in both cases a pole of order @xmath190 has the largest residue , but it is apparent that , while for @xmath187 the other poles are close to that one , in the case of @xmath186 there seems to remain a jump between the first pole and the others . this fact is in agreement with the remark , already made by stieltjes , that the poles of the laplace transform of @xmath187 become dense on the imaginary axis and corresponds with the fact that the spectral measure is in this case analytic . on the other hand , the same fact seems to suggest a qualitatively different behaviour of @xmath188 , and in particular that @xmath100 is not analytic in this case , so that the decay of the time autocorrelation of @xmath171 is not exponential . as a final remark , we add a few considerations concerning the actual difficulty occurring in the calculations . in view of theorem [ teor : stieltjes ] , all poles lie on the imaginary axis and all the determinants ( [ eq : def_determinanti ] ) are positive . so , we are certain that the outcome of our computation ceases to be reliable at the order at which we encounter a negative determinant . since the coefficients are determined via a monte carlo approximation of an integral , they are subject to a numerical error . the difficulty is that increasing the order requires also a corresponding increase of the precision , and so a longer and longer calculation . this makes the whole procedure unpractical even at not too high orders , and for this reason we performed most calculations up to order 8 only , thus finding 4 frequencies . we have shown that there is a convergent algorithm which , in principle , could be implemented to get the laplace transform of the time autocorrelation of a dynamical variable , and so to get conclusions about the rate of its decay as @xmath82 . in our opinion , the most relevant aspect is the connection we found between the spectral measure and the solution of the stieltjes moment problem , which implies a relation between the smoothness property of the measure and the fulfilment of an infinite set of constraints ( see , as an example , proposition [ prop : hausdorff ] in appendix [ app : hausdorff ] ) . this leads us to the question , whether smooth measures are in some sense exceptional , namely , to ask whether these constraints can still be satisfied when small changes in the moments are performed ( i.e. , by generic small perturbations of the hamiltonian flow ) . implementing such an algorithm on a computer raises major difficulties with respect to both the computational time and the reliability of the results . as a matter of fact , the main problem is connected with the calculation of large determinants . however , we tried to apply the method to two functions of interest fro the fpu problem , namely , the energy of a packet of low frequency modes and the kinetic energy of half of a fpu chain . for the first function , our results seem to support the conjecture that the time autocorrelations exhibit a sub - exponential decay , while no definite conclusions can be drawn for the second quantity . we remark that this is a known characteristic for chains of fpu type , namely that quantities related to the particles typically exhibit a behaviour coherent with the predictions of statistical mechanics , while this often does not happen with quantities related to the modes . on the other hand , we should emphasize that the existence of functions with a slow decay of the autocorrelations is in itself an obstacle for the ergodic behaviour of a dynamical system . for similar results , see , e.g. , @xcite . the problem of determining a necessary and sufficient condition in order that the solution of an hamburger moment problem admit a bounded positive derivative has been solved by khz and krein ( see @xcite ) . unfortunately , their criterion can not be applied to study the derivative itself . thus , we report the criterion of akhiezer and krein for checking the existence of a derivative for the hamburger moment problem for the measure @xmath86 , then we give an alternative procedure for studying the derivatives of @xmath108 , is analytic if and only if @xmath108 is ( see corollary [ cor : analitico ] ) . ] at least in the case in which the time autocorrelation is analytic on the whole real axis . we convert the problem into an hausdorff moment problem ( i.e. , a moment problem on a finite interval ) , in order to get condition on the regularity of the successive derivatives . to state the result of @xcite of interest to us , let us introduce the definition of a non - negative definite sequence of moments @xmath76 , by requiring that @xmath191 for any @xmath75 and any sequence @xmath192 of complex numbers . defined in ( [ eq : def_determinanti ] ) are all non - negative , but the converse is not true . ] then one has [ prop : ahiezer ] in order that @xmath107 exist and be bounded , it is necessary and sufficient that there exist @xmath193 such that the sequence @xmath194 defined by and @xmath195 . ] @xmath196 is non - negative definite . for the study of the successive derivatives , we observe that in paper @xcite hausdorff found a solution method for the moment problem on the interval @xmath197 $ ] , which also allows to give a criterion to check whether , at any order , the derivatives of the measure exist and are bounded . such results can be also applied to our problem if the time autocorrelation is analytic on the whole real axis , because , through an analytic change of variable , one can map @xmath197 $ ] on @xmath102 and obtain the next proposition [ prop : hausdorff ] , which provides necessary and sufficient conditions in order that the @xmath75-th derivative of @xmath108 exist and be bounded . for its statement , we need to define the auxiliary moments @xmath198 and @xmath199 : we need to preliminarily define @xmath200 then we set @xmath201 these quantities are well defined , since it is known that @xmath202 is analytic in @xmath203 and the binomial expansion converges . then , we define the quantities @xmath204 and state the following [ prop : hausdorff ] let @xmath34 be an analytic function of @xmath96 , for @xmath205 , and let @xmath109 be the measure which solves the related stieltjes moment problem . a bounded derivative of order @xmath75 of @xmath108 exists if and only there exist @xmath206 such that , for any @xmath207 , @xmath208 and @xmath209 , one has @xmath210 * proof . * the proof is performed through the euler transformation @xmath211 . this enables one to define the positive borel measure @xmath212 on @xmath197 $ ] , by @xmath213 . the relation between the moments of @xmath214 on the interval @xmath197 $ ] and the moments @xmath76 of @xmath109 can be obtained by the following chain of equalities , for @xmath215 , @xmath216 as can be seen by expanding @xmath217 via the binomial formula . it is easy to see that this implies that the coefficients @xmath198 defined by ( [ eq : momenti_hausdorff ] ) are the moments of @xmath214 . in order to study the derivatives of @xmath218 , we pass to the the moments given by the distribution function @xmath219 . this function exists if @xmath122 is analytic on the whole real axis , because in such a case its spectral measure @xmath100 tends exponentially fast towards its limit for @xmath220 and , in consequence , @xmath212 tends towards the same limit as @xmath221 at least as fast as @xmath222 . moreover , @xmath219 admits a bounded derivative of order @xmath75 if and only if @xmath223 does . we need , thus , to find the moments of @xmath219 , which are given by @xmath224 in view of the binomial expansion of @xmath225 . since this expansion converges uniformly for @xmath226 $ ] , one can exchange the order of summation and integration and prove that these moments are indeed the @xmath227 defined by ( [ eq : momenti_hausdorff ] ) . as hausdorff proved , in order that the moment problem for the sequence @xmath227 have a solution @xmath228 on @xmath197 $ ] , which is almost everywhere differentiable with a limited derivative ( in our case , it is @xmath229 ) , it is necessary and sufficient that @xmath230 be bounded , where @xmath231 are defined by equation ( [ eq : definizioni_hausdorff ] ) . this corresponds to @xmath232 then , one considers the hausdorff moment problem in which the measure ( not necessarily positive ) is @xmath233 , and expresses its moments @xmath234 through the original moments @xmath227 , by integrating by parts . so one obtains that the coefficients @xmath235 defined by ( [ eq : definizioni_hausdorff ] ) play for @xmath233 the same role which the @xmath236 play for @xmath228 , because , for @xmath237 , and @xmath238 , since they give indications only on the discontinuities of @xmath233 , which can be disposed of , since @xmath233 can be defined at will on a set of measure zero . ] @xmath239 thus , one can express the condition that @xmath233 admits a limited derivative by asking that @xmath240 have an upper bound . the proof is then concluded by iterating this way of proceeding . q.e.d . 1ex finally , we add the following proposition , which can be proved by the easy remark that a power series with positive coefficients , which has a radius of convergence @xmath241 , has a singularity for @xmath242 ( this is known as vivanti s theorem ) . [ prop : convergente ] in order that the decay of the time autocorrelation of the dynamical variable @xmath0 be not exponentially fast it is sufficient that there exists @xmath243 such that @xmath244{c_n}\le\rho$ ] . * remark 5 . * the condition of proposition [ prop : convergente ] requires only to have an upper bound on the @xmath76 , whereas those of propositions [ prop : ahiezer]-[prop : hausdorff ] ask for a more detailed knowledge on the coefficients @xmath76 . we believe that the first one is seldom fulfilled ( an example in which it is fulfilled is that of a harmonic oscillator ) . for , the @xmath75-th derivative of a function usually grows as @xmath245 , so that @xmath246{c_n}=+\infty\ .\ ] ] the other requirements enables us to deal with a larger class of variables , but , as just said , it has the drawback of needing a very detailed knowledge of all coefficients . consider the series ( [ eq : serie_troncata ] ) , and recall that the remainder has a definite sign . due to the inequalities ( [ eq : correlazione_e_varianza ] ) we readily get the bounds @xmath247 which hold true because the remainder is negative in the first case and positive in the second one . setting @xmath248 , we get the inequalities @xmath249 i.e. , the polynomials @xmath250 and @xmath251 are positive semi - definite for @xmath252 . look now for the real roots of the equations @xmath253 and @xmath254 , with @xmath255 . since @xmath250 and @xmath251 are polynomials of degree @xmath167 in which @xmath167 changes of sign occur , descartes rule of signs implies that such roots , if they exist , must be positive . thus @xmath256 and @xmath257 are positive semi - definite on the whole real axis , i.e. , the inequalities ( [ eq : richiesta_coefficienti_p ] ) hold true for all @xmath258 . consider now the dynamical variable @xmath259 , calculated by recursive application of the poisson bracket with @xmath3 . by a straightforward application of theorem [ teor : serie ] we get that the coefficients of @xmath260 are precisely @xmath261 as defined in ( [ eq : coefficienti_generalizzati ] ) . since the argument above applies to any dynamical variable , then we get that the corresponding polynomials @xmath77 and @xmath78 are positive semi - definite , too , as claimed . let us recall the fa di bruno s formula ( see @xcite ) . let a function @xmath262 be given , where @xmath0 and @xmath19 possess a sufficient number of derivatives . then one has @xmath263 where the sum is over all @xmath75-tuples of nonnegative integers @xmath264 satisfying the constraint @xmath265 in our case , we apply twice the formula , the first time to find the successive derivatives of the canonical coordinates @xmath162 and @xmath266 with respect to time , the second one to find the derivatives of @xmath267 and @xmath268 . whereas the latter application is trivial , we remark that the former is obtained by explicitly expressing the force @xmath269 as a function of @xmath162 and @xmath161 , then by writing @xmath270 as the term at the r.h.s . involves only derivatives of @xmath162 up to order @xmath271 , the scheme can be applied iteratively to obtain all the derivatives of @xmath162 ( and @xmath161 ) with respect to time . for positive integers @xmath272 we introduce the sets @xmath273 where @xmath274 is the set of non negative integers . for @xmath181 we have @xmath275 , i.e. , just one element . for @xmath276 the following statement is obviously true : _ every element @xmath277 can be written as @xmath278 where @xmath279 . _ this is the key of our algorithm . in @xmath280 we consider the inverse lexicographic order , right to left . more precisely , the ordering is recursively defined as follows : for @xmath181 the ordering is trivial , since there is just one element ; for @xmath281 with @xmath276 we say that @xmath282 if either case applies : ( i ) @xmath283 , or ( ii ) @xmath284 and @xmath285 . e.g. , setting @xmath286 and @xmath287 we get the ordered set @xmath288 remark that the last vector @xmath289 is @xmath290 . the first vector is easily found by setting @xmath291 for a given @xmath47 the next vector is found as follows : starting from @xmath292 find the first @xmath293 which is not zero , so that @xmath294 and @xmath295 . then replace @xmath293 with @xmath296 and @xmath297 with the first vector of @xmath298 . the algorithm stops when @xmath299 , namely the last vector , because the index @xmath300 can not be found . c. liverani , ann . math . * 159 * , ( 2004 ) 12751312 . n. chernov , r. markarian , _ chaotic billiards _ , ams ( providence , 2006 ) . a. carati , j. stat . phys . * 128 * , ( 2007 ) 10571077 . a.m. maiocchi , a. carati , commun . phys . * 297 * , ( 2010 ) 427445 . a. carati , a.m. maiocchi , commun . phys . * 314 * , ( 2012 ) 129161 . koopman , proc . usa * 17 * , ( 1931 ) 315318 . n. obrekov , _ verteilung und berechnung der nullstellen reeller polynome _ , veb deutscher verlag der wissenschaften ( berlin , 1963 ) . d. ruelle , phys . . lett . * 56 * , ( 1986 ) 405407 . t. j. stieltjes , ann . toulouse sr . 1 * 8 * , ( 1894 ) j1j122 and * 9 * , ( 1895 ) a5a47 . widder , _ the laplace transform _ , princeton university press ( princeton , 1946 ) . w. feller , _ an introduction to probability theory and its applications _ , vol . ii . , wiley & sons ( new york , 1971 ) . akhiezer , _ the classical moment problem and some related questions in analysis _ , oliver & boyd ( edinburgh , 1965 ) . f. hausdorff , math . z. * 16 * , ( 1923 ) 220248 . e. fermi , j. pasta , s. ulam , in _ e. fermi collected papers _ , vol . 977988 , the university chicago press ( chicago , 1965 ) . a. carati , l. galgani , a. giorgilli , s. paleari , phys . e * 76 * , ( 2007 ) 022104 . a. giorgilli , s. paleari , t. penati , extensive adiabatic invariants for nonlinear chains , preprint . n.i . akhiezer , m.g . krein , _ some questions in the theory of moment _ , ams ( providence , 1962 ) . f. fa di bruno , ann . sci . mat . * 6 * , ( 1855 ) 479480 . french version in : note sur une nouvelle formule de calcul diffrentiel , quart . j. pure appl . math . * 1 * , ( 1857 ) 359360 .	we present here a general iterative formula which gives a ( formal ) series expansion for the time autocorrelation of smooth dynamical variables , for all hamiltonian systems endowed with an invariant measure . we add some criteria , theoretical in nature , which enable one to decide whether the decay of the correlations is exponentially fast or not . one of these criteria is implemented numerically for the case of the fermi - pasta - ulam system , and we find indications which might suggest a sub - exponential decay of the time autocorrelation of a relevant dynamical variable .

Subsets and Splits

No community queries yet

The top public SQL queries from the community will appear here once available.